CLASSIFICATION OF YOUTUBE COMMENT SPAM USING TF-IDF AND MULTINOMIAL NAÏVE BAYES CLASSIFIER

Abstract

Nowadays, social networking such as Facebook, YouTube, Telegram, Instagram, etc. are very popular among people as IT technologies are developing more and more. Facebook and YouTube are the most popular social media platforms between young and old people, especially young people. The users can subscribe and give their opinion as the comments on YouTube. This system is developed with YouTube comment spam classification framework by using term frequency- inverse document frequency (TF-IDF) and Multinomial Naïve Bayes. TF-IDF is a statistical method to measure the weight or score of each word in each document to the whole corpus. This system is implemented using ASP.Net programming language on Microsoft Visual Studio 2015 IDE and Microsoft SQL Server 2017 Express Version as Database Engine. In this system, 1965 comments typed for five music videos of five singers (PSY, Katy Perry, LMFAO, Eminem, and Shakira) uploaded on YouTube are collected as data set. The purpose of this system is to categorize the YouTube comments into the suitable categories by using Multinomial Naïve Bayes Classifier and classify the comments as spam or legitimate (ham) depending on the contents in comment. Finally, the system evaluates the results with the accuracy (precision, recall and F-measure).