Volume 18, No. 6, 2021
An Iterative Approach For Processing The Large Scale Datasets
Dr. S. CHITRA
Abstract
With the introduction of Web 2.0, the amount of emotive content available on the Internet has increased. Movie or product reviews, user comments, testimonials, remarks in discussion forums, and other forms of such content are frequently seen on social networking websites. The benefits of timely discovery of emotive or opinionated web content are several, the most important of which is monetization. Understanding the feelings of the general public toward various entities and products allows for better contextual advertising, recommendation systems, and market trend analysis. The goal of this study is to develop a sentiment-focused web crawling framework that will make it easier to find and analyse emotive material in movie and hotel reviews. Statistical methods are utilised in this study to capture aspects of subjective style and sentence polarity. The research compares and contrasts the overall accuracy, precisions, and recall values of two supervised machine learning algorithms: K-Nearest Neighbour(K-NN) and Nave Bayes'. It was discovered that while Nave Bayes performed significantly better than K-NN in terms of movie reviews, these algorithms performed similarly poorly in terms of hotel evaluations..
Pages: 5284-5300
Keywords: Iterative Classification, K-Nearest Neighbour, Naïve Bayes, Web Content Mining, Sentiment analysis