COMPARISON OF SVM, KNN, AND NAIVE BAYES METHOD WITH N-GRAM IN TRAFFIC ACCIDENT CLASSIFICATION

Abstract

Traffic accidents that occur in Indonesia are still relatively high, the information can 
be easily obtained through social media, one of which is Twitter. The amount of traffic 
accident information can be processed and classified according to certain categories. 
Traffic accident data classification is done using SVM, KNN and Naïve Bayes methods 
using n-gram feature extraction. The results of this study indicate the best accuracy is 
87.63 using the KNN method.