Using text mining to handle unstructured data in semiconductor manufacturing — Yan-Hsiu Liu

In the field of semiconductor manufacturing, people usually focus on the data of well-designed databases, such as values from tool sensors, inline metrology data, or WAT data. These data are well structured and easily handled by engineers for further analysis. For example, integration engineers can compared CD metrology data to WAT data to catch out the root cause of abnormal device current, or equipment engineers can check FDC data to judge PM success or not.