Data Mining And Quality Improvement For Key Process Yield Of D Company

This thesis explores how a brand company manages its supplierquality under big-data era.In this thesis, we based on D Company`s Key Process Yield (KPY)data-base to discuss the result of quality improvement by using datamining models in the brand business-to-supplier relationships. First, webuilt D Company`s KPY data-base and created a standard KPY yield datacollection process, then we used CHAID decision tree classificationmethod to find the key factors which affect KPY, and analyzed KPY`sdistribution characteristics through box plot. We defined KPY`s categoryand benchmark rules, finally established KPY industry benchmark. Welook for quality improvement opportunities by building a supplierbenchmark yield to KPY industry benchmark, and compare it to asupplier`s historical KPY. Then we make a decision wheather there is aquality improvement opportunity. After quality improvementopportunities are identified. D Company and the supplier will together seta quality improvement project. Using the DMAIC methodology to achieve quality improvement and using Kappa test the consistency, theauthor participated in all of the above activities, and participated in theimplementation of Company D`s supplier T’s quality improvementprojects.The improvement of the quality through data mining models outperformsthe traditional point-point improvement method in D Company`s, whouses data mining and data analysis to look to quality improvementopportunities at the macro level, and then set up a project team at themicro level to implement the quality improvement campaign.Establishing industry benchmarks also introduce a competitionmechanism for quality improvement among suppliers. It made the qualityimprovement more comprehensive and accurate. This qualityimprovement strategy based on data mining models can widely be used inthe same industry and other similar industries.