[2C17-23] Fukushima 3

5:40 PM - 5:55 PM JST(8:40 AM - 8:55 AM UTC)

[2C21] Investigation on distribution of radioactive substances in Fukushima(16) Development of a prediction model for ambient dose equivalent rate distribution based on environmental half-life profiles using LASSO

*Yoshiaki Shikaze¹, Kimiaki Saito¹, Satoshi Mikami¹, Masahiko Machida¹, Kazuya Yoshimura¹ (1. JAEA)

Keywords:

predictive model,environmental half-life,machine learning,LASSO,ambient dose equivalent rate,land use,KURAMA,integrated map

We are currently developing a model that predicts the trend of changes in the ambient dose equivalent rate in the environment using the environmental half-life of multiple components, based on an environmental half-life profile in which the weights of each decay component obtained by LASSO, a type of machine learning, are normalized and averaged for each group of the ambient dose equivalent rate range of the first measurement after the accident and land use type. Here the accuracy of this model was evaluated in comparison with the integrated map.