Pencarian Fitur Optimal Halaman Web Menggunakan Kombinasi Algoritme Genetika dan Naive Bayes untuk Klasifikasi

Abstrak

The process of classification of web pages has problems in the selection of relevant features that affect the acquisition of value accuracy. These problems can be handled using feature selection techniques. Feature selection method works by evaluating and selecting relevant and informative features of each web page document. Features are a token or informative words that often appear on web pages. In this study, the method used is the genetic algorithm and decision tree incorporated in the wrapper technique. Genetic algorithms are used as subset selection and decision tree as attribute evaluators. The proposed method is able to reduce features by 45.39% for the WebKB dataset and 56.71% for the r8 dataset. Overall, the results of the classification process increased although not too significant.