Open AccessImplemented PSO-NBC and PSO-SVM to Help Determine Status of Volcanoes
Author(s) -
Firman Tempola
Publication year - 2019
Publication title -
jurnal penelitian pos dan informatika
Language(s) - English
Resource type - Journals
eISSN - 2476-9266
pISSN - 2088-9402
DOI - 10.17933/jppi.v9i2.198
Subject(s) - particle swarm optimization , naive bayes classifier , support vector machine , computer science , artificial intelligence , machine learning , classifier (uml) , bayes' theorem , data mining , pattern recognition (psychology) , bayesian probability
This research is a continuation of previous research that applied the Naive Bayes classifier algorithm to predict the status of volcanoes in Indonesia based on seismic factors. There are five attributes used in predicting the status of volcanoes, namely the status of the normal, standby and alerts. The results Showed the accuracy of the resulted prediction was only 79.31%, or fell into fair classification. To overcome these weaknesses and in order to increase accuracy, optimization is done by giving criteria or attribute weights using particle swarm optimization. This research compared the optimization of Naive Bayes algorithm to vector machine support using particle swarm optimization. The research found improvement on system after application of PSO-NBC to that of 91.3 % and 92.86% after applying PSO-SVM.