Industrial Quality Prediction System through Data Mining Algorithm

Abstract

Based on an assessment of production capabilities, manufacturing sectors' core competency is increased. The importance of product quality in this aspect cannot be overstated. Several academics have introduced Deming's 14 principles, Shewhart cycle, total quality management, and other approaches to decrease the external failure costs and enhance product yield rates. Analysis of industrial data and process monitoring is becoming increasingly important as a part of the Industry 4.0 paradigm. In order to reduce the internal failure cost and inspection overhead, quality control (QC) schemes are utilized by industries. The final product quality has an interactive and cumulative effect of various parameters like operators and equipment in multistage manufacturing processes (MMP). In other cases, the final product is inspected in a single workstation with QC. It's challenging to do a cause analysis in MMP whenever a failure occurs. Several industries are looking for the optimal quality prediction model in order to achieve flawless production. The majority of current approaches solely handles single-stage manufacturing and is inadequate in dealing with MMP quality concerns. To overcome this issue, this paper proposes an industrial quality prediction system with a combination of multiple Program Component Analysis (PCA) and Decision Stump (DS) algorithm for MMP quality prediction. A SECOM (SEmiCOnductor Manufacturing) dataset is used for verification and validation of the proposed model. Based on the findings, it is clear that this model is capable of performing accurate classification and prediction in the field of industrial quality.