Big Data and Predictive Analytics Methods for Modeling and Analysis of Semiconductor Manufacturing Processes

摘要

Semiconductor manufacturing fabs generate huge amount of data. The big data approaches of data management have increased speed, quality and accessibility of the data. This paper discusses harnessing value from this data using predictive analytics methods. Various aspects predictive analytics in the context of semiconductor manufacturing are discussed. The limitations of standard methods of analysis and the need to adopt robust methods of modeling and analysis are highlighted. The robust prediction modeling method is implemented on wafer sensor data resulting in improved prediction ability of wafer quality characteristics.