Abstract
During semiconductor fabrication process, huge process data will be automatically or semi-automatically recorded and accumulated in database for monitoring the process, diagnosing faults and managing manufacturing. However, the manufacturing factors that affect the wafer yield are frequently interrelated. Domain engineers cannot easily find possible root causes of low yield rapidly and efficiently only using their own domain knowledge or applying rules of thumb. This study aims to construct a data mining framework for analyzing semiconductor manufacturing data and propose a hybrid decision tree approach that involves Kruskal-Wallis test, chi-square interaction detection, and the variance reduction splitting criterion to analyze huge multi-dimensional data and infer possible causes of faults for troubleshooting. The proposed hybrid decision tree approach can also eliminate the variable selection bias during the decision tree construction. We conduct an empirical study in a semiconductor company for validation. The results demonstrated the practical viability of the proposed method to help the engineers to diagnose the faults and improve yield efficiently and effectively.