Intelligent Wafer Sensor Fault Detection Using Machine Learning

For electrical equipment to function properly, semiconductor wafer quality and dependability are essential. Traditional fault detection techniques frequently miss small flaws that could cause serious product failures. This paper presents an intelligent wafer sensor fault detection system using machine learning to improve reliability in semiconductor manufacturing [7]. Wafer sensor datasets are collected from Kaggle and preprocessed for noise removal, normalization, and feature extraction [3]. A Random Forest/XGBoost algorithm is applied to train the model, ensuring high accuracy and robustness in detecting faulty wafers [1], [8]. To enhance transparency and reliability, multiple datasets are utilized during training and evaluation [5]. The finalized model is integrated into a Flask-based web application with Python backend, enabling users to upload wafer sensor data and receive real-time fault predictions. This system aims to reduce manual inspection, minimize production downtime, and provide a scalable solution for efficient fault detection in wafer processing [4].