Host: The Japan Society of Mechanical Engineers
Name : [in Japanese]
Date : September 11, 2022 - September 14, 2022
Model-driven approaches that use numerical simulations with observation data as initial input parameters have been mainly used in forecasting weather phenomena. On the other hand, in the recent AI boom, research on data-driven approaches that replace some of these approaches with machine learning has been actively conducted. In this presentation, I will discuss the integration of numerical simulation and machine learning in weather forecasting, and also introduce the activities to improve the accuracy of typhoon forecasting through holding a data analysis competition.
