Research on quality defect analysis and prediction model based on association rule mining

Abstract

A quality defect analysis and prediction method based on association rule mining is proposed to address the coupling and ambiguity between multiple quality data in the process of product manufacturing quality control and diagnosis. It overcomes the shortcomings of the traditional quality defect analysis method which can only trace the quality from a single chain and can simultaneously analyze and predict the specific quality characteristics data that lead to the output quality defects and the multiple input parameters of the manufacturing process that have an impact on it. By dividing the quality characteristics data intervals through K-means and using the Apriori algorithm to explore the correlation between the quality characteristics data, we can construct the rules to judge the loss of product quality. A GA-SVR based manufacturing process quality defect prediction model is built using the cloud server plus local terminal technology structure. Finally, through example analysis, it is proved the effectiveness of the proposed method.