Six Sigma and Data Mining to Improve Production Processes

Purpose – Six Sigma includes traditional tools that are usually difficult to give a full understanding of the manufacturing data. Data mining is usefully applied in industries; some algorithms generate significant information to serve the industrial fields. Data mining algorithms enhanced DMAIC to extract relevant knowledge about the process.
Design/methodology/approach – In the defining step, three weeks were spent outlining the main problems by studying and evaluating the job shop processes. Data were collected during the measuring step to evaluate the baseline process and build a reliable plan to make the change in the process. In analyzing, two data mining algorithms were applied to analyze and select the most important predictors. Worker efficiency was evaluated carefully to be applied as output in the dataset; 12 technical variables were carefully screened and filtered from huge datasets to extract significant relationships. In the improving step, the results have been applied to make the changes by adding lean principles and agile projects. Finally, five months waited to get the impact on business, so this was confirmed by the management in the control step.
Findings – The productivity of the weaving department has improved from 58% to 92%, lead time has been reduced from an average of 35 days to an average of 20 days, and the capability process has improved from 0.77 to 1.7.
Originality/value – The value lies in the systematic approach to problem-solving, the use of statistical methods to analyze data, and the application of scientific principles to improve the processes.

Six Sigma, Data Mining, Worker Efficiency, Features Selection