[O-11] Senior high school student poster presentations

[O11-P26] Prediction of Upper Wind Using Machine Learning for Launch of Flying Object

*Ritsuka Shoyama¹ (1.Shibuya Kyoiku Gakuen Makuhari Senior High School)

Keywords:Upper wind, Prediction, Machine learning, Flying object

Upper wind over Kagoshima City was predicted using machine learning based on past weather data to improve trajectory prediction accuracy in rockets and balloons' launch planning. First, one year of publicly available meteorological data was formatted for compatibility with the machine learning library. 88% of the upper wind data was randomly selected for training, and the remaining data was used for testing. Using a linear model or Ridge regression, predictions were made for the following days based on 1.5 consecutive days of input data, revealing that after 15 days, errors in wind direction and speed increased significantly. Then, as the length of the input data was extended, prediction accuracy improved with longer input periods, up to 30 days of input data.