Optimization Naive Bayes using Particle Swarm Optimization in Volcanic Activities

Abstract

This study is a continuation of previous studies that apply Naive Bayes classifier algorithm for predicting the status of volcanoes in Indonesia based on factors of seismicity. There are 5 criteria used in predicting the status of the mountain, namely the status of normal, alert and standby. The results of the study showed that the system accuracy produced was only 79.31%, in other words, it was still at the stage of fair classification. To overcome these weaknesses so that accuracy increases, optimization is done by giving the weight of criteria or attributes using particle swarm optimization. From the results of research by applying the same data using Particle Swarm optimization methods optimization, the accuracy of the resulting system increase of to 95.65%, where the number of particles is initialized 50 and the weight range [0 2].