[18p-E303-1~11] 13.1 Fundamental properties, surface and interface, and simulations of Si related materials

△ [18p-E303-6] Modeling of electron transmission process in the channel of nanoscale device under random impurity fluctuation with machine learning

〇(B)Ryuho Nakaya¹, Koudai Hirata¹, Souma Kawahara¹, Tota Suko², Masakazu Muraguchi¹ (1.Hokkaido Univ. of Science, 2.Waseda Univ.)

Keywords:semiconductor, machine learning, electron dynamics

A machine learning method for a model of the electron transmission process in the channel of nanoscale semiconductor devices under the impurity fluctuation is studied. We examined methods such as multiple regression analysis and random forest using various feature quantities considering physical properties of potential field. As a result, the highest prediction accuracy was obtained in the random forest approach when the channel region was divided by the certain area and the count the presence or absence of the impurity in that area as the feature value.