Leveraging Text Mining Techniques to Disclose Material Errors in Engineering Build Request Forms

Abstract

In the electronics industry, driven by continuous innovation, the preparation of precise engineering build request forms (EBRs) is critical for ensuring product quality and efficient production processes. Errors in these documents can lead to significant financial losses, including penalties and project cancellations. This research aims to enhance the error-checking process in EBRs by applying text-mining techniques through Python programming, integrating Optical Character Recognition (OCR) and automated text processing to improve inspection accuracy and efficiency. The methodology involves comparing the performance of human inspectors with the automated program using paired t-tests to assess performance. The results demonstrate that the automated program significantly reduces inspection time and increases accuracy, thereby minimizing production errors and associated costs. This advancement is anticipated to substantially improve the efficiency and reliability of document preparation in the production process, leading to more accurate and timely product outputs.