영향권을 활용한 태풍정보에 따른 피해정도 예측 분석

Abstract

According to the Disaster Yearbook, the damage caused by typhoon amounts to about 210 billion won per year. In particular, the occurrence period of typhoon overlaps with the harvest period of crops, causing great damage to farms. Therefore, this study aims to predict the degree of damage to typhoon that will occur in the future by analyzing the damage caused by typhoon information. In order to utilize various points of the typhoon for analysis, the concept of influential sphere was introduced and analyzed. As a result of the analysis, the best results were shown in the case where the inverse distance weighted average was used as the representative value of typhoon information in the random forest model. Through the results of variable importance, it was found that the most important variable in the model was the distance. Research should be conducted from various perspectives to provide a more accurate degree of damage in the future, and it is expected that if appropriate measures are prepared based on accurate damage prediction, many farmers will be reduced from typhoon.