Analisis Prediksi Bencana Angin Puyuh di Jawa Barat Menggunakan Algoritma K-NN dan C4.5 Berbasis PSO

Abstract

One of impact from climate change which is not realized in the country of Indonesia is increase of the whirlwind disaster each years. Whereas based on data taken from the BNPB Indonesia, incidence of whirlwind disaster significantly during the last 10 years. From 2007 that only 127 incidents for year. But in the 2017 reached 645 incidents. And Losses caused by the whirlwind disaster is not small, not only material losses but also recorded hundreds of people injured and dozens of people died. Therefore we tried to classify the weather conditions during the whirlwind during 2017 using Knn and C4.5 algorithms based PSO with case studies in west java. With the approach of weather attribute data taken from BMKG and attribute of the whirlwind incident form BNPB Indonesia, this research is able to produce the best result of accuracy 83%, precision 85%, recall 95,15 % and AUC 0.591 using K-nn algorithm based pso