Knowledge Discovery about online supplier’s performance Project report submitted In the partial fulfilment of the requirement for the degree Bachelor of Science In Mechanical Engineering By Aqib Aurangzeb

Knowledge Discovery about online supplier’s performance Project report submitted In the partial fulfilment of the requirement for the degree Bachelor of Science In Mechanical Engineering By Aqib Aurangzeb (14-ME-10) Ethisham Nazakat (14-ME-15) Hammad Daud (14-ME-21) Sohail Sadique (14-ME-72) Supervised By; Dr. Bilal Akbar DEPARTMENT OF MECHANICAL ENGINEERING MIRPUR UNIVERSITY OF SCIENCE AND TECHNOLOGY (September 2018) CERTIFICATE It is certified that work contained in the project titled “Knowledge Discovery about online supplier’s performance” by “Aqib Aurangzeb, Ethisham Nazakat, Hammad Daud and Sohail Sadique” has been carried out under my supervision and this work is original and has not been submitted anywhere else for a degree. Signature of supervisor Dr. Bilal Akbar Mechanical Engineering Department MUST, AJK September, 2018 DECLARATION We declare that this written submission represents our ideas in our own words and where other’s ideas or words have been included, we have adequately cited and reference the original sources. We also declare that we have adhered to all principles of academic honesty and integrity and have not misrepresented or fabricated or falsified any idea/data/fact/source in our submission. We understand that any violation of the above will be cause for this disciplinary action by the university and can also evoke penal action from the sources which have thus not been properly cited or from whom proper permission has not been taken when needed.

Aqib Aurangzeb (14-ME-10) _____________________ Ethisham Nazakat (14-ME-15) ______________________ Hammad Daud (14-ME-21) ______________________ Sohail Sadique (14-ME-72) ______________________ Date:___________________ APPROVEL SHEET This project report titled “Knowledge Discovery about online supplier’s performance” by Aqib Aurangzeb, Ethisham Nazakat, Hammad Daud and Sohail Sadique is approved for the degree of “Bachelor of Science In Mechanical Engineering”. Examiners ________________ ________________ ________________ Supervisor(s) ________________ ________________ ________________ Chairman ________________ ________________ ________________ Date: ________________ Place ________________ Thesis Title Knowledge Discovery about online supplier’s performance Acknowledgement All glories are only entitled to ALMIGHTY ALLAH, the creator and designer of this universe, you and us without a doubt. He is the only one who gave us the necessary zeal and vigour to continue with our seemingly impossible project. We thank our parents who have been so understanding, attentive and generous with their prayers throughout this difficult period and, of course, to all our teachers who have been a great help and source of enormous inspiration.

We wish Dr. Bilal Akbar, who was the source of inspiration to undertake this project. As project supervisor, Dr. Bilal Akbar was a constant source of innovative ideas, encouragement and guidance throughout our work. Table of contents Contents TOC o “1-3” h z u Chapter 01 PAGEREF _Toc524545699 h 11. Introduction PAGEREF _Toc524545700 h 11.1 Background: PAGEREF _Toc524545701 h 11.2 Supplier performance PAGEREF _Toc524545702 h 11.3 Knowledge Discovery and Data Mining PAGEREF _Toc524545703 h 21.4 Aims and Objectives: PAGEREF _Toc524545704 h 41.1.1Aim of Project PAGEREF _Toc524545705 h 41.1.2Objectives: PAGEREF _Toc524545706 h 41.5 Thesis Breakdown: PAGEREF _Toc524545707 h 4Chapter: 02 PAGEREF _Toc524545708 h 5Literature Review PAGEREF _Toc524545709 h 52.1 Supplier Performance PAGEREF _Toc524545710 h 52.2 Framework for supplier performance: PAGEREF _Toc524545711 h 52.3 Parameters for supplier Performance PAGEREF _Toc524545712 h 62.4 Previous Methodologies PAGEREF _Toc524545713 h 72.5 Knowledge discovery: PAGEREF _Toc524545714 h 10Chapter 03 PAGEREF _Toc524545715 h 12Research Methodology PAGEREF _Toc524545716 h 12Proposed Methodology PAGEREF _Toc524545717 h 123.1 Developing an understanding of the goal and parameter identification: PAGEREF _Toc524545718 h 123.2 Collecting a target data set PAGEREF _Toc524545719 h 123.3 Data cleaning and pre-processing.

PAGEREF _Toc524545720 h 123.4 Choosing the data-mining task. PAGEREF _Toc524545721 h 133.5 Choosing the data-mining algorithm(s) PAGEREF _Toc524545722 h 143.6 Model Building. PAGEREF _Toc524545723 h 153.7 Evaluating output and interpret mined results. PAGEREF _Toc524545724 h 153.8 Consolidating discovered knowledge.

PAGEREF _Toc524545725 h 16Chapter: 04 PAGEREF _Toc524545726 h 17Data Collection and Data preparation PAGEREF _Toc524545727 h 17Data Collection PAGEREF _Toc524545728 h 174.1 Company Overview PAGEREF _Toc524545729 h 174.1.1Legal validity PAGEREF _Toc524545730 h 174.1.2 Building Information PAGEREF _Toc524545731 h 174.2Human Resources PAGEREF _Toc524545732 h 174.2.1 Employee Headcount PAGEREF _Toc524545733 h 184.3 Current Export Situation: PAGEREF _Toc524545734 h 184.4 Quality Assurance PAGEREF _Toc524545735 h 184.4.1 Quality Control Management: PAGEREF _Toc524545736 h 184.4.2 Supplier Management: PAGEREF _Toc524545737 h 184.4.3 After Sales Service: PAGEREF _Toc524545738 h 194.5 R&D Capacity: PAGEREF _Toc524545739 h 194.5.1Current Situation: PAGEREF _Toc524545740 h 194.6 Data Analysis: PAGEREF _Toc524545741 h 214.6.1Company Overview & Export Situation: PAGEREF _Toc524545742 h 214.6.2Quality Assurance PAGEREF _Toc524545743 h 234.6.3R&D Capacity: PAGEREF _Toc524545744 h 244.7 Other Parameters/Attributes: PAGEREF _Toc524545745 h 254.8 Data Preparation PAGEREF _Toc524545746 h 284.9 Data Pre-processing PAGEREF _Toc524545747 h 284.9.1 Data integration: PAGEREF _Toc524545748 h 294.9.2 Data Reduction: PAGEREF _Toc524545749 h 294.9.3Data Cleaning: PAGEREF _Toc524545750 h 29Chapter N0: 05 PAGEREF _Toc524545751 h 31Data Analysis and Results PAGEREF _Toc524545752 h 315.1 Grade A PAGEREF _Toc524545753 h 325.2 Grade B PAGEREF _Toc524545754 h 345.3 Grade C PAGEREF _Toc524545755 h 355.4 Grade D PAGEREF _Toc524545756 h 365.5 Discussion PAGEREF _Toc524545757 h 37Conclusion PAGEREF _Toc524545758 h 39Knowledge Management PAGEREF _Toc524545759 h 39Limitations: PAGEREF _Toc524545760 h 40Future Work: PAGEREF _Toc524545761 h 40References: PAGEREF _Toc524545762 h 41 List of tables TOC h z c “Table” Table 1 Framework for supplier performance PAGEREF _Toc524548019 h 6Table 2 Prior methodologies PAGEREF _Toc524548020 h 9Table 3 Category, Variables, Parameters PAGEREF _Toc524548021 h 20Table 4 Numeric Attributes PAGEREF _Toc524548022 h 25Table 5 Parameters/attributes, data type, values PAGEREF _Toc524548023 h 25Table 6 Deleted Parameters PAGEREF _Toc524548024 h 30Table 7 Models PAGEREF _Toc524548025 h 31 List of figures TOC h z c “Figure” Figure 1 Knowledge Discovery Process PAGEREF _Toc524548044 h 11Figure 2 Company Overview & Export Situation PAGEREF _Toc524548045 h 22Figure 3 Quality Assurance PAGEREF _Toc524548046 h 23Figure 4 R&D PAGEREF _Toc524548047 h 24Figure 5 Grade A PAGEREF _Toc524548048 h 33Figure 6 Grade B PAGEREF _Toc524548049 h 34Figure 7 Grade C PAGEREF _Toc524548050 h 35Figure 8 Grade D PAGEREF _Toc524548051 h 36 Abstract: “World is drowning in data but starving for knowledge.” Knowledge discovery in databases (KDD) plays an important role in decision-making tasks by supporting end users both in exploring and understanding of very large datasets and in building models with validity over unseen data. Knowledge Discovery in Database (KDD) brings the latest research in databases and Artificial intelligence (AI). Many studies have addressed which criteria have to be considered to weight potential suppliers when traditional vendor–buyer purchasing transactions take place. However, due to the peculiarities of the e-procurement process, paradigms and criteria are studied for the selection of viable suppliers when internet-based transactions are set up. The outcomes are based, on the one hand, on the work of a focus group especially set up for this purpose, and, on the other hand, on the literature review.

Then, the proposed selection criteria are structured into an original framework. The framework is conceived for the direct in-field application by means of common multi-attribute decision-making methods, such as the data mining techniques. In this research, various algorithms, techniques and methods are used to mine data; including decision trees, rule induction. These algorithms, techniques and methods are used to detect patterns in the dataset. The first and simplest analytical step in data mining is to describe the E- suppliers data (engine part suppliers) – summarize its statistical attributes (such as means and standard deviations).

In the Data Mining Process, collecting (supplier data), exploring and selecting the right data is important. In this research, useful patterns found by using these algorithms are analysed to discover potentially useful knowledge about supplier’s performance. Chapter 011. Introduction1.1 Background:Supplier performance is becoming an increasingly important part of the solicitation process. After the emergence of high-profile scandals involving supplier misrepresentation, misappropriation of funds and outright fraud, purchasing managers across all industries are coming to understand the crucial role of research and due diligence play in the solicitation management process.

A report released by U.S. risk consulting firm Kroll, titled the Global Fraud Report, executives from major companies worldwide were polled about their concerns regarding fraud within their respective corporate structures. 42 percent of respondents identified their company as “highly or moderately vulnerable” to supplier or procurement fraud, an increase of 16 percent from 2010. Further, 20 percent of respondents said that they were in fact affected by supplier or procurement fraud in 2011. This data indicates that supplier risk is of growing concern, as real losses from supplier fraud are being absorbed by an increasing number of companies every year in the United States.

More recently, in 2013, the owner of an electrical services company based in Maryland was charged with identity fraud, identity theft and other serious crimes. It was discovered that his company failed to comply with federal wage and record-keeping regulations while also using employees’ social security numbers in an unauthorized manner. Such flagrant violations of federal and state labour laws are unfortunately all too common, however in recent years, advancements in technology and an increasing willingness on the part of suppliers and organizations to proactively address such concerns has greatly improved the climate for supplier performance initiatives. 7 1.2 Supplier performance Supplier performance is basically an audit/examine of a supplier’s processes, policies, and financial health to determine how much risk it poses to the contracting organization. 3 Assessing your suppliers is a critical aspect of your third-party performance program.

But it’s expensive, time-consuming, and often painful for both your organization, as well as each supplier. Supplier performance accomplishes this by providing the technology, process and people necessary to efficiently understand supplier risk, help you remediate inefficient controls and better prepare and protect your organization from third party risks. 4 Supplier Performance are very important as they form an integral part of an occupational health and safety management plan. They help to create awareness of hazards and risk. Identify who may be at risk. If a supplier is hypercritical to the successful production of your good or successful delivery of your service, or if it is inherently extra risky (such as a third party that handles customer payment info necessary to sell the good or service but also loaded with risk), then it’s a prime candidate for a supplier performance.

The process can audit many different areas, including adherence to industry certifications and standards, the financial health of the supplier, geographic and geopolitical factors, a supplier’s contingency plans in case of emergency (so that it still delivers its product to you), and cybersecurity. A supplier performance evaluates a multitude of factors and provides actionable information that you can use to mitigate that risk. 5 Supply chain management is a very complex set of operations and functions with an enormous range of inherent risks. These can be a minor irritation such as small delay which does not cause a significant consequence or a major problem such as fire in a supplier’s warehouse which can cause the disruption of the entire chain. Risk on the supply chain consists in every risk that might affect the planned flow of material.

6 1.3 Knowledge Discovery and Data MiningKnowledge discovery is the nontrivial process of identifying valid, novel, potentially useful and ultimately understandable patterns in data. Knowledge discovery and Data mining in databases have been attracting a significant amount of research and industry. Knowledge discovery is revolving around the investigation and creation of knowledge, processes, algorithms and the mechanisms for retrieving potential knowledge form data collection. Related issues include data collection, database design and description of entities in database using most appropriate representation and data quality. An important component of activity is identifying patterns and trends which suggest entity relationships. 2 Across a wide variety of fields, data are being collected and accumulated at a dramatic pace.

There is an urgent need for a new generation of computational theories and tools to assist humans in extracting useful information (knowledge) from the rapidly growing volumes of digital data. These theories and tools are the subject of the emerging field of knowledge discovery in databases (KDD)cross a wide variety of fields, data are being collected and accumulated at a dramatic pace. The KDD field is concerned with the development of methods and techniques for making sense of data. The basic problem addressed by the KDD process is one of mapping low-level data (which are typically too voluminous to understand and digest easily) into other forms that might be more compact (for example, a short report), more abstract (for example, a descriptive approximation or model of the process that generated the data), or more useful (for example, a predictive model for estimating the value of future cases). At the core of the process is the application of specific data-mining methods for pattern discovery and extraction.

A search engine (e.g. Google) receives hundreds of millions of queries every day. Each query can be viewed as a transaction where the user describes her or his information need. What novel and useful knowledge can a search engine learn from such a huge collection of queries collected from users over time? Interestingly, some patterns found in user search queries can disclose invaluable knowledge that cannot be obtained by reading individual data items alone.

For example, Google’s Flu Trends uses specific search terms as indicators of flu activity. It found a close relationship between the number of people who search for flu-related information and the number of people who have flu symptoms. A pattern emerges when all the search queries related to flu are aggregated. Using aggregated Google search data, Flu Trends can estimate flu activity up to two weeks faster than traditional systems can. 30 Well -known examples of Data Mining and Analytics come from E-commerce sites. Many E-commerce companies use Data Mining and Business Intelligence to offer cross-sells and up-sells through their websites.

One of the most famous of these is, of course, Amazon, who use sophisticated mining techniques to drive their, ‘People who viewed that product, also liked this’ functionality and identify customer behaviour buying pattern. 31 With computerised banking everywhere huge amount of data is supposed to be generated with new transactions. Data mining can contribute to solving business problems in banking and finance by finding patterns, causalities, and correlations in business information and market prices that are not immediately apparent to managers because the volume data is too large or is generated too quickly to screen by experts. The managers may find these information for better segmenting, targeting, acquiring, retaining and maintaining a profitable customer. 32 Extracting information and knowledge from huge volumes of raw data, data mining can assist decision-making and strategic planning in areas such as logistics and SCM. In the case of SCM, the knowledge generated could benefit the whole supply chain.

With the growing popularity of the Internet and E-Commerce, all the parties of Supply Chain are on the solid cooperation in information sharing, to achieve higher customer satisfaction and lower cost. However, integrating the data and processes among the partners might cause the information quality issues, which will influence the operational process performance significantly. Taiwan Uncle Sam Apparel company is one of the wholesalers and retailers for apparel goods. To reduce stock, the decision makers want to understand the relationships between the supply and sales.

That is, a decision tree for association rules is necessary for decision support. 33 Historically, the notion of finding useful patterns in data has been given a variety of names, including data mining, knowledge extraction, information discovery, information harvesting, data archaeology, and data pattern processing. The term data mining has mostly been used by statisticians, data analysts, and the management information systems (MIS) communities. Data mining is the application of specific algorithms for extracting patterns from data.

Data mining is a step in the KDD process that consists of applying data analysis and discovery algorithms that produce enumeration of patterns (or models) over the data. 1 In our research, we use KDD & DM techniques for extracting useful knowledge for risk assessment of e-suppliers supply chain. 1.4 Aims and Objectives:Aim of ProjectDiscovering knowledge of E-suppliers performance by recognising and evaluating risks using data mining technique. Objectives:Literature review and identification of parameters Data collection and pre-processing Model Building Data analysis and results 1.5 Thesis Breakdown:Chapter 2 is the Literature Review of previous methodologies used for supplier risk assessment and understand the application domain of knowledge discovery and data mining. Chapter 3 is the Proposed Methodology in our research. This chapter includes sequence of steps to complete research.

Chapter 4 is Data collection and data preparation. This chapter is all about the identification of parameters, collection of data related to parameters and all about the data preparation for data mining tool. Chapter 6 is Data Analysis and results. In this chapter we interpret the mined data to evaluate results or extract knowledge.

Chapter: 02 Literature Review2.1 Supplier PerformanceSupplier performance measure is a practice that is used to measure, analyse, and manage the supplier’s performance to cut costs, alleviate risks, and drive continuous improvement. The ultimate intent is to identify potential issues and their root causes so that they can be resolved to everyone’s benefit as early as possible. 35 We constantly strive to create and manage a highly reliable, competitive supply chain.To compete in today’s demanding marketplace, suppliers must be recognized leaders in our chosen markets, providing responsive, quality solutions to improve customer’s competitiveness. To measure supplier progress towards this quest, the Supplier Performance System was developed. 36 When a supplier fails to meet delivery, quality, and price requirements, additional costs are incurred by the buying organization to correct these deficiencies.

These excess costs, both direct and overhead, have an immediate impact on the firm’s available resources. Waste of people, equipment, and time, which all cost money, adversely affects the firm’s competitive position. Unsatisfactory supplier performance which results in actual additional costs often is not reflected in the purchase price or charged back to the supplier. These costs are, in fact, hidden costs. Management of the supply base requires a formal, cost-based method of evaluating a supplier’s performance. 6 2.2 Framework for supplier performance: There are the 5 key steps common among success supplier performance.

Step 1: The first step is to capture key performance metrics within the supplier’s contracts. This ensures that all key termsmeasures to ensure contract compliance are visible. Secondly, gather input from key relationship managers to understand their supplier performance objectives and use the information to establish metrics and ensure that they are aligned with overall strategy. These metrics and targets should be shared with suppliers and mutually agreed to, so both the company and suppliers can create the right performance management program Step 2: On a consistent and frequent basis, the company should collect information to calculate current values on an agreed upon set of metrics, thresholds and targets. Various methods that can be used to gather this data include supplier assessment surveys, information from ERP systems, homegrown operational systems, instant supplier feedback etc.

Step 3: Once data is collected, it should be aggregated to report on performance versus plan. While spreadsheets and other tools can be used for analysis, supplier performance management systems significantly improve the ability to analyse the information. For example, KPI’s (key performance indicators) allow companies to monitor the progress of their suppliers ensuring they get early warnings if suppliers are underperforming. KPI scores can be compared with contract terms to ensure contract compliance.

Scorecards further aggregate this information and provide companies the ability to view supplier performance at a moment in time or monitor trends over a certain period. They allow the purchasing organization to compare the performance of a supplier to those of their peers. Alerts let companies know when their suppliers are operating outside pre-established tolerances. Color-coded status buttons quickly flag potential areas for the company to focus on. Step 4: Scorecards, trend reports and alerts help identify gaps between target and actual performance for every supplier. The purchasing organization should use this information to review the impact of performance gaps on their business to prioritize them and then communicate the priorities of the gaps that need to be addressed with the supplier and ask for a remediation plan.

The use of collaborative supplier portals that provides this information to suppliers, along with the ability to set priorities ensures that nothing falls between the cracks and both parties are on the same page with respect to what is working well and what needs improvement. Step 5: Supplier performance management is not a onetime process. Performance should be tracked on an ongoing basis – both to ensure that previously identified gaps were remediated and to keep the focus on continuous benchmarking and improvement. 41 Table SEQ Table * ARABIC 1 Framework for supplier performanceFramework for supplier performance Step 1 Identify metrics, thresholds and targets Step 2 Collect data through various mechanisms Step 3 View and analyse aggregated information Step 4 Identify gaps, prioritize and communicate Step 5 Implement continuous tracking 2.3 Parameters for supplier PerformanceThe parameters relate to individual performance characteristics are delivery, quality, and service. Delivery performance parameters should be established for “timeliness of delivery” and “delivered quantity.” Any allowable deviations to the committed date and quantity are identified. An item might be considered on time.

Desired quality performance characteristics also need to be identified, established material specifications can be used as the performance guideline. Service parameters for service factor can also be established. These may include a supplier’s ability to resolve problems or his willingness to share certain technical data with the buyer. All performance parameters clearly should be consistent with the buying firm’s operating needs.

6 Dong et al. (2001) find that just in time purchasing, which is operationalized as reduced order sizes (less inventory), shorter lead times, and the use of quality control measures, leads to significantly lower logistics costs, including the costs of purchased materials, for buying organizations. Supplier performance is positively related to cost reduction. 8 Shin et al.

(2000) find a moderately significant relationship between improved supplier quality and lead times and production and quality costs for the buying organization. 9 In general, procurement costs, performance history, product and service quality, on timedelivery and environmental impact are the key criteria of supplier performance. 10 The total order cycle time, called order to delivery cycle time, refers to the time elapsed in between the receipt of customer order until the delivery of finished goods to the customer. The reduction in order cycle time leads to reduction in supply chain response time, and as such is an important performance measure and source of competitive advantage, it directly interacts with customer service in determining competitiveness. 11 An increase in delivery performance is possible through a reduction in lead time attributes.

Another important aspect of delivery performance is on-time delivery. On-time delivery reflects whether perfect delivery has taken place or otherwise and is also a measure of customer service level. 12 2.4 Previous MethodologiesWu et al. (2006) developed an analytic hierarchy process (AHP) based supplier risk assessment tool to determine the relative weights of individual risk factors. Using these weights and the probability of each risk factor occurring for a supplier, an overall risk index was computed.

Methods for assessing risk are also contained in the growing literature bases on supplier assessment and selection. 13 Michalski et al. (2000) supply risk assessment process is the comprehensive outsource risk evaluation (CORE) system, which is a tool developed by Microsoft and Arthur Anderson to prevent supply problems from arising. This assessment process identifies 19 risk factors categorized into four families. 14 Jennifer et al (2008) uses multi-criteria scoring procedure is used to calculate all risk assessment scores. For each part, the risk assessment score for each subcategory is found by first multiplying each supplier’s rating on that subcategory by the percent of production purchased from that supplier to get the individual supplier score on that part subcategory.

To calculate scores for each subcategory for a part by the weight assigned to that subcategory and then summing the resulting products within that category. Then supplier’s rating on a subcategory is first multiplied by the percentage of that part provided by the supplier, these products are summed across all parts supplied by that supplier, and the resulting sum is divided by the sum of all the percentages of parts supplied. 15 Bromiley et al, (2015) ERM (Enterprise risk management) generally attempts to consider all risks within an organization in an aggregate manner and to consider impacts, mitigation strategies, and response options with a focus on the enterprise. Although ERM is becoming increasingly popular in most business areas, it is still considered to be in its infancy, and most of the existing academic research is targeted towards finance and accounting.

16 Faisal Aqlan et al. (2015) proposes a framework for rapid assessment of risks in integrated supply chains by combining qualitative and quantitative techniques taking into consideration risk correlations and uncertainties. Use fuzzy logic to convert the qualitative value into quantitative value to calculate risk score, risk likelihood and risk impact, the proposed framework helps decision makers assess the risks per product type and compare and prioritize the risks of the different product types. Second, it develops a software application that helps the risk management decisions to be fast and easy.

The software tool is flexible, and it allows the user to add or delete agents, connect to database, and connect to simulation models. 17 Jian Liu et al. (2000) uses DEA (Data envelopment analysis) application to apply systematic analysis to decision making to consideration such as reduce number of supplier and to provide improvement target for supplier. It based on a strategic purchasing objective to reduce output variables include supplier variety and quality, and input variables include price indexes, delivery performance and distance factor. 18 Moss and Kurty et al. (1983) calculate the reliability analysis of preliminary design of Tension Leg Platform (TLP) using FMEA and FTA.

All possible failures and their impacts are identified and examined using FMEA. FTA is constructed based on the cause/impact correlation identified in FMEA. The FTA systematically describes all causes of undesired events leading to the failure mode 19 Geum et al. (2009) used FTA to describe the customers selection to proposed Service Tree Analysis (STA) which of service tree construction, qualitative analysis and quantitative analysis. The weakness of this method is the subjectivity 20 Pitt et al. (1994) applied HAZOP in manufacturing for safety assessment by defining how dangers can happen and considering their severity.

Controlling the probability of failure and severity of failure will help the safety measures cost effective. 21 Stiff et al. (2003) described the differences between spread mooring and turret mooring systems using HAZID that come out into scenario categories. The quantitative risk assessment is calculated using the structural reliability analysis between spread mooring and turret mooring. 22 Roy et al.

(2003) studied the quantitative risk assessment for storage and purification section of a titanium sponge production facility using FETI, HAZOP and FTA. FETI and HAZOP were used to find the most hazardous section in the entire plant which is Titanium tetrachloride (TiCl4), FTA is used as probabilistic analysis to describe the root cause of an events. 23 Dianous & Fievez et al. (2005) built methodology for risk assessment in industry using bow tie diagrams to identify the major accidents and the barriers.

To assess the number and the reliability of the safety functions risk graph is used so that a good risk control can be reached. 24 Jacinto & Silva et al. (2009) applied the bow tie method in large shipyard. Firstly, it was used to initial qualitative analysis and secondly to calculate the semi quantitative assessment. The accident risk level and acceptance criteria were carried out using scoring system 25 Petruska et al.

(2009) considered quantitative risk assessment approach to evaluate the risk of moored MODU in deep-water facilities. Identifying and defining the potential mooring failure and the consequences using HAZID which then used to develop the event trees of each scenario. ETA evaluates all the possible failure sequence and identification of their respective consequence 26 Ghodrati et al. (2007) modified the ETA to calculate the associated risks (i.e.

risk of shortage of spare parts) in estimation of the required number of spare parts due to not considering the characteristics of system operating environment. 27 Deacon et al. (2010) evaluated the risk of human error during offshore emergency musters with HAZOP and bow tie analysis. HAZOP is used to record failure modes, potential consequence and safeguards. Bow tie model allow analysing the human factors consequence. This method can decrease the gap between real and perceived risk in emergency preparedness if used appropriately.

28 Cockshott et al. (2005) constructed methodology for a new hazardous chemical marine terminal using probability bow ties and expanded with rapid risk ranking (RRR). The bow tie and RRR combination is called probability bow tie (PBT). 29 Constantin Blome et al. (2014) used the partial least squares (PLS) approach using SmartPLS 2.0 to simultaneously assess the measurement instrument and the hypothesised model. This approach was considered appropriate for two reasons: PLS delivers valid results even for small sample sizes and the estimates of the individual path coefficients are more conservative than in covariance-based techniques.

The relationship between green procurement and supplier performance is fully mediated by green supplier development. it shown that legitimacy concerns drive basic green procurement, whereas top management is decisive for advanced practices, such as green supplier development. 42 Table SEQ Table * ARABIC 2 Prior methodologiesApproaches Authors Specific Areas FMEA – FTA Moss et al (1983) Reliability Analysis of TLP FTA Geum et all (2009) Service Process Selection HAZOP Pitt (1994) Safety Assessment HAZID – StructuralReliability Analysis Stiff et atl (2003) Comparative Risk Analysis of Mooring FETI-HAZOP-FTA Roy et al (2003) Quantitative Risk Assessment in Production Facility FTA – ETA Jacinto & Silva (2009) Ship Building Industry HAZID – ETA Petruska et al (2009) Mooring MODU Risk Assessments ETA Ghodrati et al (2007) Spare part selection HAZOP – FTA –ETA Deacon et al (2010) Risk Analysis in Offshore Emergencies PLS Constantin Blome et al. (2014) Supplier Performance Lau et al.

(2009) use prototype process mining system for supporting knowledge discovery in a supply chain network is proposed and developed. This methodology of the proposed algorithm has been evaluated in a case study and the algorithm shows its potential to figure out the primary factors that have a great effect upon the satisfaction of the end customer in a supply. 2.5 Knowledge discovery:Data Cleaning: Data cleaning is defined as removal of noisy and irrelevant data from collection. Cleaning in case of Missing values. Cleaning noisy data, where noise is a random or variance error. Cleaning with Data discrepancy detection and Data transformation tools.

Data Integration: Data integration is defined as heterogeneous data from multiple sources combined in a common source(Datawarehouse). Data integration using Data Migration tools. Data integration using Data Synchronization tools. Data integration using ETL(Extract-Load-Transformation) process. Data Selection: Data selection is defined as the process where data relevant to the analysis is decided and retrieved from the data collection. Data selection using Neural network.

Data selection using Decision Trees. Data selection using Naive Bayes. Data selection using Clustering, Regression, etc. Data Transformation: Data Transformation is defined as the process of transforming data into appropriate form required by mining procedure.

Data Transformation is a two-step process: Data Mapping: Assigning elements from source base to destination to capture transformations. Code generation: Creation of the actual transformation program. Data Mining: Data mining is defined as clever techniques that are applied to extract patterns potentially useful. Transforms task relevant data into patterns. Decides purpose of model using classification or characterization.

Pattern Evaluation: Pattern Evaluation is defined as as identifying strictly increasing patterns representing knowledge based on given measures. Find interestingness score of each pattern. Uses summarization and Visualization to make data understandable by user. Knowledge representation: Knowledge representation is defined as technique which utilizes visualization tools to represent data mining results. Generate reports. Generate tables.

Generate discriminant rules, classification rules, characterization rules, etc. Figure SEQ Figure * ARABIC 1 Knowledge Discovery ProcessChapter 03 Research MethodologyIn this research knowledge is discovered about online supplier performance from the data obtained in certified assessment reports available on Alibaba. For this study proposed methodology is as follows. Proposed MethodologyThis proposed methodology is focused on knowledge discovery form the available data to aid the manager in procurement decisions. The proposed methodology is consisting of following steps as shown in figure 3.1.

3.1 Developing an understanding of the goal and parameter identification:In this study research, first step is to understand the goal of Knowledge discovery about the online supplier performance, however the performance is dependent variable which is affected by different parameters. Here a useful information extract about online suppliers using rule-based classification algorithms which is helpful in decision making. To achieve the goal, detail literature review is conducted to identify the important parameters which impact on the performance of suppliers (given in chapter 4). 3.2 Collecting a target data setIn this research, all the required data is collected from assessment reports of suppliers which are available online, certified assessment reports related to supplier which are providing engine parts to their customers are used as part of the data.

Certified assessment supplier reports are studied, the parameters which judge the supplier performance are selected based on the understanding the hypothesis of research and design of research which directly impact on the performance of suppliers. 3.3 Data cleaning and pre-processing.In this research, data pre-processing techniques are used to fill the missing values, identify and remove the outliers, smooth noisy data and resolve inconsistencies from the data. Data pre-processing techniques include Data integration Data reduction Data cleaning In data integration, most important and common parameters from the assessment reports are selected and plotted on Excel Spread sheet to minimise the complication in reading the data. The next step after data integration is data reduction. In data reduction the selected common and important parameters which are lengthy, and complex are reduced into short and simplest possible form to reduce the complexity of data on Excel spread sheet.

Options of attributes in parameters are converted into Alphabetical form, which then further clarifies the data and resulting the patterns to easier to understand. The last step in processing the targeted data in this research is data cleaning. In this step, the missing values of parameter filled manually by comparing the supplier data with other supplier or by using most probable value to fill the missing value. Outliers and noisy values are identified using histogram and replaced by the mean value of the attribute to decrease the variance in the data, to increase the data effectiveness and reliability (see section 4.9).

Data cleaning routines work to clean the data by filling the missing values, smooth the noisy values, identifying, removing the outliers and resolving the inconsistencies and remove variance from data by decreasing standard deviation of data. 3.4 Choosing the data-mining task.Data Mining Descriptive Data mining Predictive data mining Clustering Association Regression Prediction Classification Data mining is divided into two types: Predictive data mining. Descriptive data mining. In this research, descriptive data mining (classification) technique is used to classify the data. Rule-based classification algorithms are used to classify our data and to find patterns in the dataset and evaluate knowledge. 3.5 Choosing the data-mining algorithm(s)JRIP Part J48 Rule Induction Decision Tree Misc.

Bayes Meta Classification Figure 3.1 For classification of data, rule-based classifications algorithm and techniques in WEKA are: Rule Induction (JRIP, PART) Decision Tree(J48) Bayes Misc. meta In this research Rule Induction and Decision Tree classification algorithms are used to classify the data. Rule induction is an area of machine learning in which formal rules are extracted from a set of observations. The rules extracted may represent a full scientific model of the data, or merely represent local patterns in the data. Rule induction algorithms “Ripple” and “PART” are used. Decision tree is a decision support tool that uses a tree-like graph or model of decisions and their possible consequences, including chance event outcomes, resource costs, and utility.

3.6 Model Building.In this research, JRIP (ripple), PART and J48 (decision tree) algorithms are used to build different models by using WEKA as a data mining tool. 10 folds cross validation testing method is used to mine the data for rule-based classification algorithms. The different attributes are deleted in building different models and properties of algorithms are also changed to enhance the accuracy of model. Batch size (The preferred number of instances to process) ranges from 50-200, folds (determine amount of data used from pruning) ranges from 1- 10, min No ( the minimum total weight of instances in a rule) ranges from 1- 6, optimization (the number of optimizations runs) is ranges from 1-10, seed (used for randomizing the data) ranges from 0-2, confidence factor (used for pruning in which smaller value incur more pruning) ranges from 0.125-1.0. JRIP Algorithm: This class implements a propositional rule learner, Repeated Incremental Pruning to Produce Error Reduction (RIPPER), which was proposed by William W. It deals with Binary attributes, Date attributes, Empty nominal attributes, Missing values, Nominal attributes, Numeric attributes, Unary attribute.

PART Algorithm: Class for generating a PART decision list. Uses separate-and-conquer. Builds a partial C4.5 decision tree in each iteration and makes the “best” leaf into a rule. J48 Algorithm: Class of generating a pruned or unpruned C4, J48 is an extension of ID3 (Iterative Dichotomiser 3).

The additional features of J48 are accounting for missing values, decision trees pruning, continuous attribute value ranges, derivation of rules, etc. In the WEKA data mining tool, J48 is an open source Java implementation of the C4.5 algorithm (C4.5 builds decision trees from a set of training data in the same way as ID3, using the concept of information entropy. At each node of the tree, C4.5chooses the attribute of the data that most effectively splits its set of samples into subsets enriched in one class or the other). In this study research, JRIP (ripple) algorithm is chosen as it has the best accuracy than PART and decision tree algorithms.

3.7 Evaluating output and interpret mined results.In this research, different patterns with respect to the supplier’s grades which will assign to them are evaluated. Also change the properties of algorithms to make output better and then interpret these patterns to evaluate knowledge about the supplier performance. 3.8 Consolidating discovered knowledge. This research incorporates the knowledge into the performance system, or simply documenting it and reporting it to users. Now data is in a whole unit or in the form of information which can be implemented.

Proposed Methodology for Supplier performance . Developing understanding of goal and parameter identification Evaluating output and interpret results Consolidating discovered knowledge Collecting target data set for supplier’s performance Data cleaning and pre-processing Choosing the data-mining task Choosing the data-mining algorithms Model building Evaluating output and interpret mined results Consolidating discovered knowledge Figure 3.1 Chapter: 04 Data Collection and Data preparation Data CollectionIn this research, data is collected from the assessment reports of certified assessed supplier of Alibaba E-commerce. Based on the understanding the hypothesis of research and design of research and to gathered information about variables, first, assessment reports of almost 2000 certified supplier’s reports have been downloaded and identify the variables and parameters which play an important role in supplies risk assessment and on the behalf of performance of the suppliers can be judged. Detail of Categories, variables and parameters which are used in this research for judgement of E-suppliers is given below 4.1 Company OverviewA company overview (also known as company information or a company summary) is an essential part of a business plan. It’s an overview of the most important points about your company your history, management team, location, mission statement and legal structure. The variables in this category are: Legal validityLegal validity means company includes valid business license, its experience, type of ownership and company address etc.

The parameters of Legal validity are: Validity period Valid business license Industry Experience Export Experience Annual review conducted by Industrial Bureau 4.1.2 Building InformationBuilding information is about its building certification, its size and number of buildings. The parameters of Building information are: Total Building Size Number of Buildings Human ResourcesA human-resources department (HR department) of an organization performs human resource management, overseeing various aspects of employment, such as compliance with labour law and employment standards, administration of employee benefits, and some aspects of recruitment and dismissal. The variable in this category is: 4.2.1 Employee HeadcountEmployee headcount means department name, part and full-time employees. The parameter in this variable is: No of employee 4.3 Current Export Situation:Export situation is about foreign trading employees in the company, valid export license of company and trade agents’ employee overseas of the company. The Parameters for Current export situation are: Foreign trading Employee Valid export license Trade agents employed overseas Average Lead time 4.4 Quality AssuranceQuality assurance (QA) is a way of preventing mistakes and defects in manufactured products and avoiding problems when delivering solutions or services to customers.

Major variables in Quality assurance are: 4.4.1 Quality Control Management:The act of overseeing all activities and tasks needed to maintain a desired level of excellence. This process may include implementing a strategy for quality planning, quality assurance, quality improvement and quality control. The main parameters of quality control management are: Is there is quality control on production line? Does QA/QC work independently from Production line? How many QC inspectors in total? 4.4.2 Supplier Management:The process of obtaining and managing of products or services needed to operate a business or other type of organization. The purpose of supply management procedures is to keep costs stable and use resources effectively to increase the profits and efficiency of the business or organization. The Parameters of supplier management are: Does company have supplier assessment Procedure? Does Company have updated list of approved suppliers? Does company keep its supplier assessment report? Is there random product inspection after final packaging? 4.4.3 After Sales Service:After sales service refers to various processes which make sure customers are satisfied with the products and services of the organization. The needs and demands of the customers must be fulfilled for them to spread a positive word of mouth.

The parameters of this variable include: Is customer feedback including complaints are clearly recorded? Is there is closed loop corrective system is placed? Is there a product alert or recall procedure? 4.5 R&D Capacity:Research and development (R&D) is a valuable tool for growing and improving your business. R&D involves researching your market and your customer needs and developing new and improved products and services to fit these needs. The variable in R&D capacity are: Current Situation:In R&D situation, the number of R&D engineers in company, third party testing of products and standard design procedures for designing is considered. The parameters of current situation variable are: R&D Engineers in company Are there relevant design input/output, review, and verification documents available for the assessment company? Do R& D employees use any specific software for designing new products? Has the company established standard design procedures for new products? Have the designed products been tested by third-party inspection body? Are the designed products confirmed by the customers? Does the company have qualification requirements for designers? In this research, the parameters/attributes which are random or whose value remain same for all suppliers are deleted. Deleted parameters are: Validity period Is there is quality control on production line? Does QA/QC work independently from Production line? Table SEQ Table * ARABIC 3 Category, Variables, ParametersCategory Variables Attributes/Parameters Definition Company Overview Legal validity Export experience An increase in the cumulative value exported of a product, or to a foreign market Industry experience Industry experience is one of the first requirements in distinctive asset for a professional Ownership type Ownership of property may be private, collective, and the property may be objects, land/real estate, or intellectual property Building Information Certification Type It is a designation earned by a person to assure qualification to perform a job or task.

Total Building Size A structure occupies a certain portion or land parcel. Human Resources Employee Headcount No of Employee Employee is individual, hired by an employer to do a specific job. Current Export Situation Current Export Situation Foreign trading employee Who exchange the capital, goods and services across intentional borders or territories. valid export license Export control document issued by a government agency to monitor the export Employed Overseas A person that sells another company goods or products abroad.

Average lead time The time between the initiation and completion of a production process. Quality Assurance Quality Control Management QA/QC inspectors Quality control inspectors examine products and materials for defects or deviations from manufacturer or industry specifications. Supplier Management Supplier assessment procedure Company having supplier assessment procedure Company updated list Company having updated list of approved suppliers Supplier assessment reports Company Keeps the assessment reports of supplier. Product inspection after packaging Company having a procedure for product inspection after final packaging. After Sales Service Customer feedback Customer feedback e.g.

complaints about products are recorded. Closed-loop corrective system Ensure that everyone involving in solving problem is aware of its status. Product recall Return of product, after discovery of safety issues. R&D Capacity Current Situation R&D engineers R&D engineers and tasked with directly developing new products and tasked with applied research in scientific or technological fields, which may facilitate future. Relevant design input/output, review and verification Relevant design input/output, review, and verification documents available for the assessment company. Specific software for designing Company having specific software’s for designing.

Standard design procedures To design products with generally accepted and uniform procedures, dimensions or materials Products tested by third-party Independent organization reviewed the manufacturing process of product and determined that product is comply with standards. Products confirmed by customers Company ensure customer start interacting with product right way. Qualification requirements Fulfilment of necessary conditions such as training schooling. Table 4.1 In this research, data about above parameters/attributes in table is collected from the assessed report of certifies suppliers of Alibaba E-commerce on the Excel spread sheet.

4.6 Data Analysis:In this research two types of data analysis are used which will help in modelling and transforming data with the goal of discovering knowledge and supporting decision making. Company Overview & Export Situation:A company overview (also known as company information or a company summary) is an essential part of a business plan. It’s an overview of the most important points about your company your history, management team, location, mission statement and legal structure. Export situation is about foreign trading employees in the company, valid export license of company and trade agents’ employee overseas of the company.

Figure SEQ Figure * ARABIC 2 Company Overview & Export SituationIn this research above graph is about No. of supplier verses type of ownership and certification type of company. In 2012 suppliers, 1823 suppliers have private ownership, 90% suppliers have private ownership while 119(7%) suppliers have sole proprietorship and the remaining 69(3%) have joint venture ownership. When Certification of suppliers is judged it is seen that 971 suppliers have lease type certification which is 48.2%. while 607 suppliers have factory claimed certification. A few suppliers (195) have real state certification and remaining have land certification.

From this graph it is seen that private ownership has a great impact on the performance of suppliers. Lease type certification and factory office claimed has a large effect on performance of supplier. The above graph explains the how many suppliers have valid license and how many has not. This graph also explains whether the suppliers are trade agents or not. In this graph from 2012 suppliers 1731 suppliers have valid license it means about 86% suppliers have valid license it positively effect on the company performance.

The company will go ahead. In this overview trade agents have a very less percentage about 10% suppliers are trading agents it means that a few suppliers who are selling goods in abroad. It negatively impacts on their performance. Quality AssuranceQuality assurance (QA) is a way of preventing mistakes and defects in manufactured products and avoiding problems when delivering solutions or services to customers. Figure SEQ Figure * ARABIC 3 Quality AssuranceIn the above graph it is seen that supplier reports keep for previous 12 months(c) and keep for 1-3 years(b) have a great percentage progresses the company.

521 suppliers were not keeping assessment reports whereas 294 suppliers(a) are keeping assessment reports for more than 3 years. 1185 suppliers inspected Product after final packaging with clear standard and write inspection record(a). it means it has a great impact because most of supplier using inspection after final packaging, it improves the performance.1285 suppliers give feedback to customers with standard form and records. This aspect improves the supplier performance. A low percentage (4.1%) of suppliers who are not giving feedback to customers(d) .67% corrective action is taking place.

this attribute has a great importance, if it will take place then supplier performance will increase. Whereas, product recall has less percentage in this research after discovery of any fault no safety factors which badl1y impact on supplier performance. Product recall percentage is 16.4. when overview of assessment procedure is taken it is seen that 1732 suppliers have an assessment procedure which is about 86%. Product is assessed after final packaging it directly improves supplier performance. Approved supplier’s percentage greatly impact on company performance which indirectly grade up the supplier.

R&D Capacity:Research and development (R&D) is a valuable tool for growing and improving your business. R&D involves researching your market and your customer needs and developing new and improved products and services to fit these needs. Figure SEQ Figure * ARABIC 4 R&DIn the above graph it is seen that 818 suppliers have standard design procedure with written records(a) and 503 suppliers which also has standard design procedure, but they have no written record. Suppliers who has design procedure with written record, company give them more importance, it increases supplier performance. And 800 suppliers(a) have a procedure that all designed products confirmed by third party but 574 suppliers whose products are confirmed only designed part(b). only 5 suppliers whose products are not confirmed by third party. When Qualification requirements are judged it is seen that 656 suppliers need qualification requirement with job description whereas 479 suppliers have qualification requirement but without written job description.

504 suppliers have no qualification requirements. Qualification requirement is necessary to enhance the performance of suppliers. 57% relevant design I/O review is conducted but here 43% has not. Here a little performance of suppliers has increased. And when look at specific software 1394 employees use specific software for design which is about 70%.

This case enhances the performance of suppliers which is profitable to the company. 4.7 Other Parameters/Attributes:These attributes are the numeric attributes which are tabulated below. A numeric attribute is quantitative; that is, it is a measurable quantity, represented in integer or real values. Numeric attributes can be interval-scaled or ratio-scaled.

Table SEQ Table * ARABIC 4 Numeric AttributesSr. No Attributes Min. Value Max. Value Mean Value 1 Export Experience 0 37 9.2 2 Industry Experience 0 37 10.4 3 No. of Employees 3 221 64.3 4 Foreign Trading Employee 0 20 7.2 5 Average Lead Time 0 60 22.4 6 QA/QC Inspectors 0 12 3.29 7 R&D Engineers 0 14 3.28 In the above table it is seen that maximum export experience for company is 37 years.

And the mean experience for company 9 years and 2 months. But the mean industry experience is 10 years and 5 months. Minimum no. of employees required for company is 3, as no.

of employees increases company makes progresses more. And average number of employees for the company are 64. Cycle time which is the most important parameter for supplier performance, Maximum 60 days whereas the mean/Avg. cycle time among 2000 suppliers is 22 days.

Max. Quality assurance inspectors for company are 12 and Research and development engineers are 14. With increasing number performance will get better but extra cost must pay. So that’s why here company hire QA/QC and R&D engineers according to the required work. Table SEQ Table * ARABIC 5 Parameters/attributes, data type, valuesAttribute Data Type Values Attribute Data Type Values 1. Export Experience Numeric 0-37 Years 16.Customer feedback Nominal Yes, with a standard feedback form and records 2.Industry Experience Numeric 0-37 Years Yes, with a standard feedback form but no records 3.Ownership Type Nominal 1.Private Owner Yes, with records but no standard feedback form 2.Public Company No 3.Joint Venture 17.Close loop corrective system Nominal Yes 4.Stated Owned No 5.Sole Proprietors 18.Recall procedure Nominal Yes 6.Other No 4.Certification Type Nominal 1.Land Certification 19.R&D engineers Numeric 0-14 2.Real Estate Certification 20.Relevant design input/output Nominal Yes 3.Lease Agreement No 4.Factory Officer Claimed 21.Specific software for designing Nominal Yes 5.Building size Numeric 40-881 m2 No 6.No Of employee Numeric 3-221 22.Standard design procedures Nominal Yes, with clear written instructions 7.Foreign Employee Numeric 0-20 Yes, without written instructions 8.Valid Export License Nominal Nominal Yes No No 23.Products tested by third-party Nominal Yes, all designed products have been tested 9.Employed overseas Yes Yes, only part of designed products has been tested No No 10.Lead time Numeric 0-60 Days 24.Products confirmed by customers Nominal Yes, all designed products have been confirmed 11.Q/A inspectors Numeric 0-12 Yes, part of designed products has been confirmedaccording to client’s requirements 12.Assessment Procedure Nominal Yes No No 13.Updated Supplier Nominal Yes 25.Qualificat-ion requirements Nominal Yes, with written job description No Yes, without written job description 14.Keep assessment report Nominal Yes, assessment reports are available for more than 3years No, but at least two years design experience isneeded Yes, assessment reports are available for the last 1-3years No Yes, assessment reports are available for the previous12 months Grade Nominal A No B 15.Product Inspection Nominal Yes, with clear standard and written inspection records C Yes, with inspection records but no procedures D Yes, with procedures but no inspection records No, inspections are not necessary 4.8 Data PreparationData Pre-processing technique is carried out to prepare the data for data mining.

Real world data is often incomplete, so there is need to resolve issues and transforming such data into an understandable format. Data pre-processing is an important step in the data mining process. Data-gathering methods are often loosely controlled, resulting in out-of-range values, impossible data combinations. Pre-processing includes fill in missing values, smooth noisy data, identify or remove outliers, resolve inconsistencies, part of data reduction, replacing and transformation of voluminous data into simplified form.  4.9 Data Pre-processing Data pre-processing technique includes: Data Integration Data Reduction Data Cleaning Data Transformation 4.9.1 Data integration: In this research, after the selection of targeted parameters/attributes which influence the performance of supplier, the data from 2000 assessed supplier reports is integrated on excel worksheet. 4.9.2 Data Reduction: The attributes “Validity period, is there is quality control on production line? Does QA/QC work independently from Production line?” which are random for all the supplier are deleted because they cannot influence the mined results. The attributes/parameters “Does company keep its supplier assessment reports?, Is there random product inspection after final packaging?, Is customer feedback including complaints are clearly recorded?, Has the company established standard design procedures for new products?, Have the designed products been tested by third-party inspection body?, Are the designed products confirmed by the customers?, Does the company have qualification requirements for designers?” had large values, are replaced by the alphabetical notation to get simplified form of dataset so that the resulting mining process may be more efficient, and the patterns found may be easier to understand.

Data Cleaning:In this research, the dataset has most of suppliers in dataset have missing value of attribute “R;D Engineers” the missing value are filled manually by comparing the data this supplier with other suppliers’ data or filled by most probable value or the mean value of attribute in dataset. There are some suppliers which have several attributes with missing values, they are ignored/deleted from dataset. Outliers and noisy data in the dataset is identified using histogram and noisy values and outliers are replaced by the mean value of attribute to remove inconsistency from dataset and to reduce the standard deviation of data that is the deviation of data from mean value. Because high standard deviation results in low effectiveness and low reliability of data that directly impact the quality of mining results.

The attributes “No of employee, Foreign trading employee, Q/A Q/C inspectors and R;D Engineers” have many noisy values that was creating variance in the data, their standard deviation is very high than mean value, these noisy values in the attributes are replaced by mean values of these attributes to decrease the standard deviation than mean and to get better mined results. The Attribute “Total Building size” which is a basically numeric attribute, but data type shown in “WEKA” is sting/ordinal nominal value, the filter “Ordinal to Numeric” is used to convert the data type of “Total building size” to numeric. And, in the attributes “Keep assessment reports, Products inspection after final packaging and Third-party inspection body” value was repeated and the filter “Merging two values” is used to merge the same values of these attributes to avoid complexity in patterns. Data after pre-processing is subject oriented, integrated, time variant and non-volatile and can be used for data mining. Table SEQ Table * ARABIC 6 Deleted ParametersModel No. Deleted Parameters Changing properties Algorithms Accuracy 1 7,18 No JRIP 66.03 Part 63.0 J48 60.6 2 9,14,20,22 No JRIP 66.6 Part 61.3 J48 60.1 3 9,21 No JRIP 66.4 Part 59.8 J48 59.4 4 10,17 No JRIP 67.4 Part 62.05 J48 60.7 5 12,14 No JRIP 67.08 Part 62.1 J48 61.4 6 25 No JRIP 68.2 Part 61.7 J48 60.1 7 2,10 Yes JRIP 69.1 Part 67.6 J48 64.6 8 12,14 Yes JRIP 70.4 Part 65.6 J48 61.4 9 25 Yes JRIP 70.06 Part 61.5 J48 58.3 10 11,19 Yes JRIP 71.4 Part 62.9 J48 60.5 Chapter N0: 05 Data Analysis and ResultsData analysis is a process of inspecting, cleansing, transforming, and modelling data with the goal of discovering useful information, informing conclusions, and supporting decision-making.

After collecting the targeted data and pre-processing the collected data and then we put the data in data mining tool (WEKA) for classification. We use rule-based classification algorithms (JRIP, PART and Decision tree) to build model of our data in data mining tool “WEKA” to get useful trends or patterns and accuracy of models are judged. We build different models using JRIP, PART and J48 algorithm, we use hit and trail method by deleting different attributes in different models and also observe the results by changing properties of algorithms. Table SEQ Table * ARABIC 7 ModelsModels Algorithms TPR FPR Precision F-measure ROC area incorrectly classified instances Correctly classified instances 1 JRIP 0.660 0.230 0.689 0.644 0.75 33.9 66.03 Part 0.630 0.168 0.626 0.637 0.76 36.9 63.0 J48 0.606 0.189 0.603 0.601 0.74 39.3 60.6 2 JRIP 0.666 0.226 0.695 0.651 0.75 33.3 66.6 Part 0.614 0.180 0.610 0.611 0.74 38.6 61.3 J48 0.602 0.184 0.599 0.600 0.74 39.8 60.1 3 JRIP 0.665 0.230 0.697 0.648 0.74 33.5 66.4 Part 0.598 0.180 0.597 0.598 0.74 40.17 59.8 J48 0.594 0.194 0.589 0.590 0.73 40.5 59.4 4 JRIP 0.675 0.232 0.719 0.658 0.76 32.5 67.4 Part 0.621 0.168 0.619 0.619 0.77 37.9 62.05 J48 0.607 0.189 0.604 0.604 0.74 39.2 60.7 5 JRIP 0.671 0.223 0.700 0.655 0.75 32.9 67.08 Part 0.622 0.178 0.617 0.618 0.75 37.8 62.1 J48 0.615 0.179 0.612 0.612 0.75 38.5 61.4 6 JRIP 0.683 0.213 0.710 0.671 0.77 31.7 68.2 Part 0.618 0.176 0.614 0.615 0.76 38.2 61.7 J48 0.601 0.183 0.599 0.600 0.74 39.8 60.1 7 JRIP 0.691 0.207 0.717 0.680 0.78 30.8 69.1 Part 0.676 0.143 0.675 0.676 0.77 32.3 67.6 J48 0.646 0.157 0.646 0.646 0.75 35.3 64.6 8 JRIP 0.705 0.177 0.714 0.698 0.79 29.5 70.4 Part 0.656 0.145 0.657 0.657 0.75 34.3 65.6 J48 0.615 0.179 0.612 0.612 0.75 38.5 61.4 9 JRIP 0.701 0.181 0.709 0.692 0.79 29.9 70.06 Part 0.616 0.172 0.614 0.615 0.74 38.4 61.5 J48 0.584 0.193 0.584 0.583 0.75 41.6 58.3 10 JRIP 0.714 0.177 0.725 0.706 0.79 28.5 71.4 Part 0.630 0.171 0.625 0.627 0.76 37.04 62.9 J48 0.605 0.173 0.606 0.605 0.77 39.4 60.5 We observe that JRIP algorithm gives better accuracy for classification of data than PART and J48 algorithm, so we consider JRIP algorithm and among the JRIP algorithm classification model, we use the model with the best accuracy. 5.1 Grade A(Industry experience <= 12.8) and (Export experience >= 7) and (Total building Size: <= 44) and (Total building Size: >= 17) and (Foreign trading employee <= 3) (35.0/4.0) (certification Type: = Lease Agreement) and (Foreign trading employee >= 7) and (customer feedback = c) and (Export experience >= 7.8) and (Employee >= 39) and (Industry experience <= 12) and (keep assessment reports = d) (22.0/0.0) (Industry experience <= 14) and (Foreign trading employee >= 6) and (average lead time <= 18) and (Industry experience >= 9) and (Export experience <= 9) and (Total building Size: <= 25) (18.0/1.0) (certification Type: = Lease Agreement) and (Standard design procedures = a) and (Foreign trading employee <= 9) and (Foreign trading employee >= 9) and (Product inspections after final packaging = a) and (Export experience <= 12) and (Export experience >= 5 (18.0/1.0) (Industry experience <= 14) and (Export experience >= 12.8) and (Product inspections after final packaging = d) and (customer feedback = b) (10.0/0.0) (Third-party inspection body = b) and (Industry experience <= 14) and (Qualification requirements = c) and (Industry experience >= 13.5) (11.0/1.0) (Industry experience <= 10.6) and (Total building Size: >= 124) and (Total building Size: <= 134) and (Foreign trading employee >= 6) (13.0/1.0) (Industry experience <= 7.3) and (Employee >= 86) and (Product inspections after final packaging = c) and (certification Type: = Lease Agreement) (9.0/1.0) (Foreign trading employee >= 6) and (Industry experience <= 8.7) and (Products confirmed = b) and (Total building Size: <= 192) and (Total building Size: >= 86) (21.0/6.0) Figure SEQ Figure * ARABIC 5 Grade AIn the above graph it is seen that industry experience has a large impact on the performance of supplier whose percentage is 88.8%. Industry experience range changes from 6-14 years.

Industry experience <= 12.8 when Export experience >= 7. Export experience changes from 7-12.8 years. Export experience is also necessary for the improvement of suppliers whose percentage is 55.5%. Total building size ranges from 17m2-192m2.

In A classification total building size impact percentage on the performance of suppliers is 44.4. Minimum foreign trading employees are 3 and maximum employees are 9. Foreign trading employees has a great impact here because they innovate product and sale company product in abroad whose percentage is 66.6. maximum supplier’s certification is lease. Certification is necessary up to 33.3% in class A.

No. of employees range from 39-86 in A class and their percentage is about 22.2%. Product is inspected with clear standard and written inspection records. Feedback to customers with standard form but no records. Products are confirmed by the customers but only the designed part. It has a little impact its percentage is 11.1.

5.2 Grade B(Total building Size: >= 588) and (Total building Size: <= 650) and (Employee >= 59) and (Foreign trading employee <= 8) (1166.0/300.0) Figure SEQ Figure * ARABIC 6 Grade BIn the above graph it is seen that total building size has a great impact on the performance of supplier whose percentage is 50. Building size area is greater than 588m2 but less than 650m2. And if no. of employees greater than or equal to 59 than supplier performance kept in grade B., but this attribute has less impact than building size about 25%.

And when seen at the foreign trading employees if less than or equal to 8 then supplier performance will have kept in Grade B. And this parameter has also less impact on performance. 5.3 Grade C(Foreign trading employee <= 4) and (Employee >= 114) and (Product inspections after final packaging = d) (16.0/2.0) (Total building Size: >= 95) and (Industry experience >= 7.8) and (Total building Size: <= 108) (36.0/8.0) (Foreign trading employee <= 4) and (Employee <= 13) and (Total building Size: <= 63) and (customer feedback = c) (19.0/1.0) (Total building Size: <= 104) and (Export experience <= 7.8) and (Product inspections after final packaging = c) and (Foreign trading employee >= 8) and (average lead time <= 15) (19.0/2.0) Figure SEQ Figure * ARABIC 7 Grade CIn the above graph it is seen that total building size and foreign trading employees have much more importance on the performance of suppliers whose percentage is 75. Total building size ranges from 63m2-108m2. And foreign trading employees ranges from 4-8.

Industry experience, Export experience and customer feedback percentage is very low, but these are the important parameters for judgement whose percentage is 25. So that’s why Grade is low. Export experience <= 7.8 and Industry experience >= 7.8 and feedback is given to customers but not with standard forms and records. Average lead time also has a very low percentage (25%) which is <=15, so that’s why kept in Grade C.

No. of employees and product inspection percentage is 50. And minimum no. of employees is 13 and maximum employees are 114.

Whereas product is inspected with procedure but not with records. 5.4 Grade D(Industry experience >= 19.6) and (Total building Size: <= 146) and (Employee <= 52) (23.0/5.0) (Product inspections after final packaging = a) and (Employee >= 100) and (Foreign trading employee <= 5) and (Foreign trading employee >= 4) and (Industry experience >= 15) (20.0/2.0) (Foreign trading employee <= 4) and (Employee <= 72) and (Employee >= 66) and (Products confirmed = a) (11.0/0.0) (Foreign trading employee <= 4) and (Total building Size: >= 616) and (Employee >= 103) (9.0/0.0) (Export experience <= 1) and (Industry experience >= 2) and (Valid export license? = Yes) (14.0/2.0) (keep assessment reports = a) and (Relevant design input/output review and verification documents = No) and (Total building Size: >= 17) and (Total building Size: <= 294) (28.0/9.0) (average lead time >= 17) and (Total building Size: >= 581) and (Total building Size: <= 628) and (Standard design procedures = a) (18.0/4.0) (Total building Size: <= 65) and (Product inspections after final packaging = c) and (Export experience >= 7.5) and (Employee <= 13) (13.0/1.0) Figure SEQ Figure * ARABIC 8 Grade DIn the above graph it is seen that total building size and number of employees impacting on the performance of suppliers and these attributes percentage is 62.5. No. of employees ranges from 13-103.

Total building size ranges from 17m2-616m2. Industry experience has low percentage (37.5%) which also enhance performance of suppliers and its ranges from2-19.6. Export experience has also very low percentage (12.5%) and it ranges from 1-7.5 years which is low experience so that’s why kept in Grade D. Cycle time which is the most important parameter for judgement of suppliers has a very low percentage (12.5%) and it is >=17. Suppliers assessment reports and product confirmed by the customers has a very low percentage (12.5%) which has much importance for supplier judgment criteria, so that’s why Grade is D. 5.5 DiscussionThe impact of industry experience on suppliers is that quality of the products improves immensely due to the time spent within the industrial zone.

The quality of the products is likely to be improved for suppliers with more industrial experience and because of having more industrial experience the supplier is likely to have an edge in being competitive which will decrease the costs of making for the suppliers. 37 However, in this research, the patterns that are obtained by using rule-based algorithms differ from the conventional methodology i.e. suppliers in the range of 6 to 14 years’ experience are found to obtain grade A. whilst suppliers with greater industrial experience are failing to obtain the top grade.

This phenomenon when researched explains that this situation is caused because suppliers with greater industrial experience have huge number of customers hence they charge more from their customers as they are cashing out their established name hence the cost is higher. Due to greater number of orders, suppliers with greater industrial experience fail to deliver on time and their quality is also decreased thus they fail to obtain the top grading. The impact of export experience on suppliers is beyond huge, because a supplier with greater export experience reflects that their business is trusted internationally by their customers resulting in the growth of supplier’s supply chain. Another impact of greater export experience for a supplier is that their export experience will diversify risk in the minds of their customers as it will reflect the years of providing trusted supply chains. Last but not the least, suppliers with greater export experience will provide better margins for their customers by decreasing the costs of their supply. 38 In this research the conventional methodology is proved by the patterns that are found from the rule-based algorithms suppliers with export experience greater than 12.8 years are given top grades whilst those failing to obtain the grade have lesser export experience.

Impact of having lease agreement certification type for a supplier is mainly concerned with the quality of the supplier’s supply chain. When a supplier is producing goods within a leased type certification faculty, they are bound to making quality products to supply to decrease the risk of the goods being rejected in the market. The would have the pressure of paying the lease on due time hence the pressure would increase the efficiency and decrease the lead times by the supplier. The patterns obtained by using rule-based algorithms prove the conventional methodology in this research. Impact of having foreign trading employees for suppliers is that the suppliers become able to target wider international markets to grow their businesses hence the quality of the supplier improves by great margins because of the feedback from the targeted markets.

Suppliers become more competitive by having more foreign trading employees and this then impacts the cost of their supplied goods as the customer will look for better margins. Having access to global markets will mean huge growth in business for the suppliers thus they will become innovative which will automatically benefit their customers. In this research, it can be seen from the patterns obtained by using rule-based algorithms that suppliers that obtain grade A, are all in possession of foreign trading employees in greater numbers. Product inspection after final packaging by the suppliers helps them improve their service as well as the quality of the supplied goods being improved. When a product is inspected before being delivered to a customer will result in winning the trust of the specific market or the customer which will then result in making the supplier more competitive in the market. Patterns obtained by using rule-based algorithms in this research show that all suppliers that manage to obtain the top grade acquire the facility of product inspection after final packaging.

Customer feedback is taken very seriously by the suppliers because this very factor helps a supplier to improve their service and quality to win customers trust. Customer feedback helps a supplier to meet the demand of the market thus they become more innovative to compete. This factor is emphasized from the patterns obtained by using the rule-based algorithms in this research. Standard design procedure is a fundamental factor for suppliers as it helps to diversify risk of their product failing in the market. Adopting to standard design procedure for a specific supply chain will enable the supplier to innovate resulting in improving the quality of their product as well as making the supplier more competitive. 39 Lead Time is an important factor for customer satisfaction.

Typically, customers want goods or service as fast as possible with minimal effort. For manufacturing and assembly, the concept of Lead Time is married to and has a direct relationship with the amount of inventory that exists at different points in the overall supply chain. Suppliers desire lesser lead time to compete in the market by winning customer satisfaction. 40 Building size is an important aspect for suppliers, having greater building size will allow the suppliers to have greater inventories and in period of emergency large orders, the suppliers become able to compete the market which diversifies the risk of losing customers. However, in this research it can be seen from the patterns obtained by using the rule-based algorithms that suppliers acquiring building size in the range of 25 to 192 square meter fall in the top-grade category meanwhile suppliers that acquire greater industrial building size fail to get grade A.

this phenomenon occurs due to the extra building size cost actually affects the cost of goods provided by those suppliers and the quality of the goods are also affected in inventories whilst suppliers that acquire lesser building size actually have faster delivery time and lesser overall cost of goods due to smaller expenses on building size. ConclusionSupplier performance has become increasingly important in today’s competitive and globally dispersed environments. The existing supplier performance models and software’s are not comprehensive and require a long time to perform risk assessment process. To compete in today’s demanding marketplace, suppliers must be recognized leaders in our chosen markets, providing responsive, quality solutions to improve customer’s competitiveness. To measure supplier progress towards this quest, the Supplier Performance System was developed.

Companies may face significant supply risk problems by overlooking supplier related threats in their supplier evaluation and selection process. Supplier evaluation and selection is one of the most critical and strategic processes in the company. Therefore, supplier related risks should be decreased by eliminating unreliable and risky suppliers. This can be achieved by considering classification -based model for supplier evaluation and selection problem. Here it was attempted to classify supplier performance by means of Rule based classification techniques (JRIP).

Classification is a well-studied area in data mining. Data mining is one step at the core of the knowledge discovery process, dealing with the extraction of patterns and relationships from large amounts of data. A rule-based classifier is a technique for classifying records using a collection of” if … then …” rules. In this research it has been discussed that the uncertain model for both numerical and categorical attributes, which are the most common types of attributes encountered in data mining. This research focuses on the uncertain attributes and assumes the class type. Suppliers divided into four classes through Rule based classification. This classification tells the researcher that which supplier is best according to the supplier selection attributes. Based on the evaluation of the classification for many supplier performance attributes, the researcher can conclude that the suppliers in class A have the highest rank. The researcher needs to consider that the supplier evaluations for different attributes and the data are examined to understand the risk structure of each supplier class and reach a more accurate interpretation of the results. Four classes of suppliers are calculated and evaluated. Class D has the highest risk scores for most of the risk types and can be labelled as suppliers with a high-risk level”. Class A has the lowest risk scores for most of the risk types related with the capabilities and commitment to business; therefore, it can be labelled as “suppliers with a low risk level”. The suppliers in class B have average risk scores; hence, Cluster B can be labelled as “suppliers with a medium risk level”. The company may prefer to work with the suppliers in Class A, or they may consider the risk criteria to make a final decision. Knowledge ManagementIn this research, the discovered knowledge is descriptive and is potentially useful for company managers as a supportive tool for taking decisions selecting suppliers according to their existing performance. Parameters such as building size and industry experience are important in this case of study as it is focusing on engine part suppliers and hence building size and industry experience of the specific supplier creates a direct impact on their performance. however, if the research target data is on Information technology industry then the impact of parameters will automatically vary as the focused industry changes. Hence the knowledge discovery about specific parameters and their impacts on different industries are managed according to the type of businesses that require the supporting tools to make their final decisions. Limitations:Precise to Chinese suppliers. Specifically, engine parts data. Limited to Rule based classification Future Work:Use large volume of data. Use different algorithms (clustering etc.) Use European and American supplier’s data. References:Usama Fayyad, Gregory Piatetsky-Shapiro, and Padhraic Smyth (1996) “From Data Mining to Knowledge Discovery in Databases” Volume 17 No. 03. Melanie J. Norton (1999) “Knowledge Discovery in Databases” volume 48 No. 01. What is supplier risk assessment?https://blog.cvmsolutions.com/what-is-a-supplier-risk-assessmentA 3rd Party Supplier Security Assessment Servicehttp://www.dvvs.co.uk/third-party-risk-assessment-managed-service/supplier risk assessmenthttps://blog.cvmsolutions.com/supplier-risk-assessment-vs.-vendor-risk-assessmentRobert M. Monczka and Steven J. Trecha “Cost Based Supplier Performance Evaluation” Journal of Purchasing and Materials Management, Spring 1988 supplier risk assessment challenges and solutionshttp://www.sourcesuite.com/procurement-learning/purchasing-articles/supplier-risk-assessment-challenges-and-solutions.jspDong, Y., Carter, C.R. and Dresner, M.E. (2001), “JIT purchasing and performance: an exploratory analysis of buyer and supplier perspectives”, Journal of Operations Management, Vol. 19 No. 4, pp. 71-83 Shin, H., Collier, D.A. and Wilson, D.D. (2000), “Supply management orientation andsupplier/buyer performance”, Journal of Operations Management, Vol. 18 No. 3, pp. 317-33. Weber C A, Current J R, and Benton W, Vendor selection criteria and methods, European Journal of Operational Research, 1991, 50(1): 2–18 Christopher, M., 1992. Logistics and Supply Chain Management. Pitman Publishing, London. Stewart, G., 1995. Supply chain performance benchmarking study reveals keys to supply chain excellence. Logistics Information Management 8 (2), 38–44. Wu, T., Blackhurst, J. and Chidambaram, V. (2006), “A model for inbound supply risk analysis”, Computers in Industry, Vol. 57 No. 4, pp. 350-65. Michalski, L. (2000), “How to identify vendor risk”, Pharmaceutical Technology, Vol. 24 No. 10, pp. 180-4. Jennifer V. Blackhurst, Kevin P. Scheibe, Danny J. Johnson, (2008),”Supplier risk assessment and monitoring for the automotive industry”, International Journal of Physical Distribution & Logistics Management, Vol. 38 Iss: 2 pp. 143 – 165 Bromiley P, McShane M, Nair A and Rustambekov E (2015). Enterprise risk management: Review, critique, and resear.ch directions. Long Range Planning 48(4): 265–276 Faisal Aqlan (2015) A Software Application for Rapid Risk Assessment in Integrated Supply Chains, Expert Systems with Applications Jian Liu, Fong-Yuen Ding, Vinod Lall, (2000),”Using data envelopment analysis to compare suppliers for supplier selection and performance improvement”, Supply Chain Management: An International Journal, Vol. 5 Iss: 3 pp. 143 – 150 H.H. Moss, M.W. Kurty, (1983). Reliability Analysis of a Tension Leg Platform. IEEE Xplore. Y. Geum, H. Seol, S. Lee, Y. Park. Application of Fault Tree Analysis to the Service Process: Service TreeAnalysis Approach. Journal of Service Management, Vol. 20 No.4, 2009 pp.433-454 M.J. Pitt. 1994. Hazard and Operability Studies: A Tool for Management Analysis. Facilities, Vol. 12 No 13. pp.5-8 J. Stiff, J. Ferrari, A. Ku, R. Spong. Comparative Risk Analysis of Two FPSO Mooring Configurations. OTC15377. 2003 P.K. Roy, A. Bhatt, C. Rajagopal. Quantitative Risk Assessment for Accidental Release of Titanium Tetrachloridein A Titanium Sponge Production Plant. 2003. Journal of Hazardous Materials A 102(2003) 167-186 V. Dianous., C. Fievez. ARAMIS Project: A More Explicit Demonstration of Risk Control Through the Use ofBow Tie Diagrams and The Evaluation of Safety Barrier Performance. Journal of Hazardous Material 130 (2006)220-233 C. Jacinto, C. Silva. A Semi Quantitative Assessment of Occupational Risks Using Bow Tie Representation.Safety Sci. (2009), doi: 10.1016/j.ssci.2009.08.008 D. Petruska, H. Banon, P. Liagre, and M. Leary, 2009. Moored MODU Risk Assessment. Offshore TechnologyConference 20165. Ghodrati, P.A. Akersten, U. Kumar. Spare Parts Estimation and Risk Assessment Conducted at Choghart IronOre Mine A Case Study. Journal of Quality in Maintenance Engineering Vol. 13 No.4, 2007. pp.353-363 T. Deacon, P.R. Amyotte, F.I. Khan. Human Error Risk Analysis in Offshore Emergencies. Safety Sci. (2010),doi: 10.1016/j.ssci.2010.02.013 J.E. Cockshott. Probability Bow Ties A Transparent Risk Management Tool. Journal of Process Safety andEnvironmental Protection, 2005. 83(B4):307-316 Jiawei Han, Micheline Kamber, Jian Pie “Data Mining Concepts and Techniques” Third Edition Chapter 01. Real life applications of Data Mining and Business Intelligence RICHARD THELWEL https://www.matillion.com/insights/5-real-life-applications-of-data-mining-and-business-intelligence/Data mining applications http://bigdata-madesimple.com/14-useful-applications-of-data-mining/Huang, S.-M.; Hu, T.-C.: Knowledge Discovery for Supply Chain Management Systems: A Schema Composition Approach. In: Issues in Information System, 5.2004, P. 523 – 529 H.C.W. Lau, G.T.S. Ho, Y. Zhao, N.S.H. Chung “Development of a process mining system for supporting knowledge discovery in a supply chain network” Int. J. Production Economics 122 (2009) pp 176–187 SUPPLIER PERFORMANCE MANAGEMENT https://www.hcmworks.com/blog/what-is-supplier-performance-managementGroupe Meloche Inc. – Strategic Suppliers Scorecard System – Rev. Oct 2011 Bob Reiss https://www.entrepreneur.com/article/205868Charles Purdy Director at Smart Currency Exchangehttps://www.businessdestinations.com/work/opinion/the-benefits-and-risks-of international-trade/Importance of Standard Procedure in Procurement of Supplies by Jeffrey Careyhttps://smallbusiness.chron.com/importance-standard-procedure-procurement-supplies-76100.htmlWhat is Lead Time, why is it important, and how do you reduce it? by Roland Lester https://www.linkedin.com/pulse/what-lead-time-why-important-how-do-you-reduce-roland-lesterftp://public.dhe.ibm.com/software/commerce/emptoris/offers/ZZW03123-USEN-00.pdfConstantin Blome, Daniel Hollos & Antony Paulraj (2014) Green procurementand green supplier development: antecedents and effects on supplier performance, International Journal of Production Research, 52:1, 32-49