Machine Learning Approach for Characteristics Prediction of 4H-Silicon Carbide NMOSFET by Process Conditions

Abstract

In this work, the electrical characteristics with various process conditions in 4H-Silicon Carbide NMOSFET are analyzed by using various machine learning algorithms. Here, the methods of machine learning, Light Gradient Boost Machine (LGBM) and Deep Neural Network (DNN) are compared for data analysis generated by Technical Computer- Aided (TCAD) simulation. Through the methods of machine learning, we analyzed the influence of process parameters for electrical characteristics. This work provides the vision that machine learning is a powerful method for analyzing and optimizing the performance of the 4H-silicon carbide MOSFET.