Volume 18, No. 6, 2021
A Study On The Sentiment Analysis Based Hybrid Collaborative Filtering Recommendation Algorithm
Analp Pathak , Dr. B. K. Sharma
Abstract
Since Internet data is growing, recommender systems must be developed and improved to provide a list of suitable favourites. Recent years have seen a rise in the popularity of review-based recommender systems, in large part due to the proliferation of social networking sites. The idea behind these kinds of systems is to put to good use the insights gained from users' written comments. The development of opinion mining and sentiment analysis has grown significantly in recent years. Hence, it led to an increase in the amount of work that has been put into the improvement of recommendation systems. The purpose of this study is to present a comparative study on sentiment analysis based on a hybrid collaborative filtering recommender system. The approach under consideration is broken down into two stages: the phase of sentiment analysis, and the phase of recommendation. The first part, "sentiment analysis," involves the estimation of sentiment scores by employing datasets collected from a variety of social media sites. The second phase involves the use of hybrid collaborative filtering. This research shows a significant increase in recommender system performance by combining sentiment analysis and other such systems. Further, we evaluated the efficacy of our hybrid collaborative filtering methods to that of alternative methods traditionally employed in sentiment analysis. From the findings, we may conclude that our suggested hybrid collaborative filtering recommendation algorithm for sentiment analysis is superior to the state-of-the-art methods already available with an accuracy of 99.2%.
Pages: 8414-8429
Keywords: Sentiment Analysis, Collaborative Filtering, Recommender Systems.