Vectorizer Comparison for Sentiment Analysis on Social Media Youtube: A Case Study

Abstract

Youtube is a popular social media used by several companies to market their products, both in the form of advertisements and videos. Nokia is one company that uses Youtube as social media to advertise and market its products until now. Nokia was a cellphone company that had fallen in 2013 due to the company's unwillingness to follow the operating system trend at the time. Nokia continues to rise and launch new products that are increasingly sophisticated. In seeing and summarizing public opinion towards the revival of the Nokia company, this research will classify the sentiment given by the public towards latest Nokia products through comments on the videos of Nokia products on Youtube. This research using Support Vector Machine (SVM) and K-Nearest Neighbor (K-NN) algorithm to classify, with comparing performance of three vectorizers, namely CountVectorizer, TFIDFVectorizer and HashingVectorizer. Compared to other algorithms and vectorizers, SVM with TFIDFVectorizer has the highest accuracy with score of 97.5%. The best vectorizer in this research is TFIDFVectorizer because there are almost no errors in predicting negative values, and also has many positive predictive values compared to other vectorizers. So, the best way to do classification is using SVM algorithm with TFIDFVectorizer.