SUPPLIER PERFORMANCE EVALUATION PREDICTIVE MODEL FOR DIRECT MATERIAL USING MACHINE LEARNING APPROACH IN SEMICONDUCTOR MANUFACTURING

Aug 3, 2026

Upload by Levi

In semiconductor manufacturing, evaluating supplier performance for direct materials is often unreliable and biased, failing to reflect suppliers’ true performance. This paper proposes a data-driven Supplier Performance Evaluation (SPE) predictive model focused on direct materials. The model applies six machine learning methods—Logistic Regression, Support Vector Machine, Naïve Bayes, Generalized Linear Model, Decision Tree, and Random Forest—to generate more unbiased supplier assessments. Results indicate Logistic Regression is best overall, with AUC-ROC of 0.993, enabling identification of material withdrawal trends. The model supports monitoring, risk management, and proactive supplier management to improve supply chain efficiency.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages15

Reading Time38 min

AccessFree to download

1.0 Introduction

Supplier Performance Evaluation value and limitations

2.0 Related Study

Traditional SPE challenges and bias

Methods such as MCDM and DEA

Related integrated and fuzzy evaluation frameworks

SUPPLIER PERFORMANCE EVALUATION PREDICTIVE MODEL FOR DIRECT MATERIAL USING MACHINE LEARNING APPROACH IN SEMICONDUCTOR MANUFACTURING S.H. Yee1,2, S.A. Asmai1*, Z. Abal Abas1, S. Ahmad1, A.S. Shibghatullah3, D. Petrovic4 1Fakulti Teknologi Maklumat dan Komunikasi, Universiti Teknikal Malaysia Melaka, Hang Tuah Jaya, 76100 Durian Tunggal, Melaka, Malaysia. 2STMicroelectronics Sdn. Bhd., Tanjong Agas Industrial Area P.O. Box 28 Muar, Johor, 84007 Malaysia. 3College of Computing & Informatics (CCI), Universiti Tenaga Nasional, 43000 Kajang, Selangor, Malaysia. 4Nottingham Business School, Nottingham Trent University, Nottingham, NG1 4FQ, United Kingdom. *Corresponding Author’s Email: [azirah@utem.edu.my](azirah@utem.edu.my) Article History: Received 7 January 2024; Revised 16 June 2024; Accepted 4 July 2024 ©2024 S.H. Yee et al. Published by Penerbit Universiti Teknikal Malaysia Melaka. This is an open article under the CC-BY-NC-ND license ([https://creativecommons.org/licenses/by-nc-nd/4.0/](https://creativecommons.org/licenses/by-nc-nd/4.0/)). ABSTRACT: In semiconductor manufacturing, evaluating supplier performance for direct materials is often unreliable and biased, failing to accurately represent suppliers' true performance. The objective of this paper is to present a data-driven Supplier Performance Evaluation (SPE) predictive model for direct material in semiconductor manufacturing. By using multiple machine learning techniques, the model provides unbiased evaluations of supplier performance. The model uses six machine learning methods: Logistic Regression, Support Vector Machine, Naïve Bayes, Generalized Linear Model, Decision Tree, and Random Forest. . The results show that Logistic Regression outperforms the other techniques with regards to analyzing both data from incoming material checks and the assembly in-process. The AUC-ROC value is 0.993 from Logistic Regression, proving that the model can identify material withdrawal trends effectively. In conclusion, the resulting model can enhance monitoring, risk management, and proactive supplier management, which leads to an efficient supply chain. KEYWORDS: Supplier Performance Evaluation; Supply Chain Management; Semiconductor Manufacturing; Machine Learning, Logistic Regression. 1.0 INTRODUCTION Supplier Performance Evaluation (SPE) is crucial to determine supplier performance with regards to complying with contract specifications of product, and service level agreements as well as achieving Key Performance Indicators (KPIs) . It can assist the organization to fulfill its objectives by establishing explicit requirements for suppliers and promoting transparency, so enabling suppliers to comprehend and pursue excellence. SPE enhances supplier quality and experience by the monitoring and evaluation of corrective actions taken by providers, including response times to complaints. This leads to a reduction in unnecessary costs through better delivery performance and tracking of product quality. Supplier Performance Evaluation (SPE) offers material providers valuable insights into operational efficiency, capacity for growth, and chances for optimization in several areas such as production schedules, technical issues, supply chains, and quality management [1] . Efficient supply chain management is crucial in semiconductor manufacture to fulfil client requirements. Although the industry has experienced substantial expansion, there is a lack of comprehensive quantitative data of SPE Conventional survey-based SPE methodologies can lead to selection bias, which can have an impact on the reputation of suppliers and the decisions made about source selection [2] . To address these challenges, the research proposes a new SPE model that functions continually, offering more unbiased and dependable data. This methodology enhances the alignment with modern data analytics, leading to a more accurate and unbiased assessment of suppliers. Paper remains arranged a

FAQs

Why is Supplier Performance Evaluation (SPE) important in semiconductor manufacturing?

SPE is crucial for assessing suppliers against contract specifications and service-level agreements while achieving KPIs. It also supports quality monitoring, corrective action evaluation, and cost reduction through better delivery performance.

What problem does the proposed predictive model address?

Conventional survey-based SPE methods can introduce selection bias and unreliable evaluations that do not represent true supplier performance. The paper addresses this by providing a continual, data-driven SPE model designed to be more unbiased and dependable.

Which machine learning method performs best and what evidence is reported?

Logistic Regression outperforms the other techniques for analyzing both incoming material checks and assembly in-process data. The paper reports an AUC-ROC value of 0.993, indicating strong capability to identify material withdrawal trends.

Aug 3, 2026

Upload by Levi