A Comparative Study on Sentiment Analysis Using KNN and SVM Models

J. Uma, Dr. K. Ramesh

Abstract: Today in this modernized era, millions and trillions of people were started sharing their day-to-day updates through the micro-blogging platforms like Twitter using where they are allowed to share their opinions and feelings within 140 characters and this restriction was done to make them concise and clear on what they are expressing through. This has turned out to be the richest source for analysis of Twitter data sentiment and belief mining by mining the sentiments both from online reviews and the social media platforms by utilizing the approach of a bag of words. Compared to other approaches, the bag of words approach is better as it accounts only for the words individually along with their count instead of neither considering the whole paragraph nor the sentence fully. Hence, this proposed paper was aimed at developing such a kind of classifier that can automatically correct and classify the unknown tweet sentiment with its label classification done by implementing few accurate techniques. Two methods are introduced in this proposed work: sentiment classification algorithm (SCA) based on support vector machine (SVM) and sentiment classification algorithm (SCA) based on k-nearest neighbour (KNN) and both methods can be evaluated for their performance depending upon the type of real tweets.

Keywords: Twitter, Twitter data, KNN, Micro-blogging platform, social media, Sentiment, Support vector machine

How to Cite?: J. Uma, Dr. K. Ramesh, "A Comparative Study on Sentiment Analysis Using KNN and SVM Models", Volume 14 Issue 12, December 2025, International Journal of Science and Research (IJSR), Pages: 1239-1244, https://www.ijsr.net/getabstract.php?paperid=SR251211124908, DOI: https://dx.doi.org/10.21275/SR251211124908