[3H01-04] Young Researcher's Session

10:30 AM - 10:45 AM JST(1:30 AM - 1:45 AM UTC)

[3H03] Feature Importance Evaluation for Core Characteristics Prediction Using Machine Learning

*Amane Haga¹, Rikuto Kasama¹, Tomohiro Endo¹, Akio Yamamoto¹ (1. Nagoya Univ.)

Keywords:

Machine Learning,Core Calculation,Fuel Loading Pattern,Neural Network,Feature Importance

Since the computational cost of core calculations is expensive, machine learning-based prediction of core characteristics has been investigated. The predictive accuracy of machine learning is significantly influenced by the input data, i.e., features. Therefore, to improve the prediction accuracy, we evaluate the feature importance in the fuel loading pattern using permutation importance.