YouTube Spam Comment Detection Using Support Vector Machine and K–Nearest Neighbor

Aqliima Aziz (Universiti Tun Hussein Onn Malaysia (UTHM))
Cik Feresa Mohd Foozy (Applied Computing Technology (ACT) Universiti Tun Hussein Onn Malaysia (UTHM))
Palaniappan Shamala (Applied Computing Technology (ACT) Universiti Tun Hussein Onn Malaysia (UTHM))
Zurinah Suradi (Dhofar University)

Indonesian Journal of Electrical Engineering and Computer Science
Vol 12, No 2: November 2018

Social networking such as YouTube, Facebook and others are very popular nowadays. The best thing about YouTube is user can subscribe also giving opinion on the comment section. However, this attract the spammer by spamming the comments on that videos. Thus, this study develop a YouTube detection framework by using Support Vector Machine (SVM) and K-Nearest Neighbor (k-NN). There are five (5) phases involved in this research such as Data Collection, Pre-processing, Feature Selection, Classification and Detection. The experiments is done by using Weka and RapidMiner. The accuracy result of SVM and KNN by using both machine learning tools show good accuracy result. Others solution to avoid spam attack is trying not to click the link on comments to avoid any problems.