Implementasi CRISP-DM Model Menggunakan Metode Decision Tree dengan Algoritma CART untuk Prediksi Curah Hujan Berpotensi Banjir
Msy Aulia Hasanah, Sopian Soim, Ade Silvia Handayani
Abstract
Indonesia is part of a tropical climate with high rainfall intensity. High rainfall intensity can potentially cause flooding. To minimize this, accurate weather predictions are needed to be able to anticipate beforehand. This research was conducted with the aim of classifying based on the rain category with the dichotomy of heavy rain and very heavy rain using data mining techniques with the CRISP-DM methodology. The algorithm used in the classification technique is CART (Classification And Regression Tree) with Confusion Matrix test parameters. Based on the results of the model evaluation, it shows that the CART algorithm has a fairly good performance in classifying with an accuracy value of 89.4%.