Finding The Available Website Name By Using Naive Bayes Classification
Kanokphon Kane · 2022 International Conference on Decision Aid Sciences and Applications (DASA) · 2022
This study aims to positive good text or negative text sentiment on websites. The property consists of the URL of the website and results in a positive website and a negative website that should be blocked in access to that. In this research, a computational approach for website classification based on features retrieved from URLs is proposed to must compare the concept of classification with the Multinomial Naive Bayes of independence property data by comparing the ideal models for discrete data Multinomial Naive Bayes. In addition, have been used other machine learnings are Decision Tree and Logistic Regression to compare which method is suitable for the analysis of this data set and which has the most accuracy with use data divided into 3 sets as follows: Training data, Test data, and External data about racist and sexist tweets sentiment to provide machine learning. The result of the Multinomial Naive Bayes has the highest accuracy of 95.14 percent for the training model.