COMPARISON OF PREDICTION ACCURACY BETWEEN DECISION TREE, NAÏVE BAYES AND K-NN ON WEB PHISING

  • Sidharta Sidharta Bina Nusantara Institute of Creative Technology
  • Albert Verasius Dian Sano Bina Nusantara Institute of Creative Technology
Keywords: Decision Tree, Naïve Bayes, K-NN, Phising

Abstract

The objective of this research is to find the best performing predictive model in term of accuracy among three classification models on web phising dataset,i.e., decision tree, naive bayes, and k-NN. The dataset has 1353 examples and 9 regular attributes and a class attribute describing whether a website is phishy or not. Cross validation with 10 folds repetition is applied to each model for training and testing. Particular parameters that significantly affect the performance are set to get optimized for each model. The result of this study shows that the best performing predictive model is decision tree model.

Published
2019-03-10