A study for big-data (Hadoop) application in semiconductor manufacturing

The semiconductor industry as one of the world's most automated and advanced manufacturing produces a huge variety of data every day. How to maximize the usage of these data is what we want to discuss in this paper. A big-data platform and its applications for semiconductor manufacturing based on Hadoop framework are proposed. Hadoop is a distributed data storage and computing solution with related low hardware cost and high system reliability. It is a great significance to conducting Hadoop platform into semiconductor manufacturing that improve effectiveness and reduce production cost. To realize the aim, we illustrate two applications of statistics methods, such as PCA (Principal Component Analysis) and SVM (Support Vector Machine). PCA can be used to optimize KPI (Key Performance Indicator) enactment through the correlation between the parameters. These KPI can provide effective monitor of process anomalies. SVM is a machine learning method is useful to build a semiconductor IKL (Intelligent Knowledge Library) to store and apply engineering lesson learns based on super large amount of data. Engineers can fulfil quick safety checking based on the mode setting by SVM.