A Novel approach of Sentiment Classification using Emoticons
Shivani Bahri, Pranav Bahri, Sangeeta Lal · Procedia Computer Science · 2018
Sentiment analysis is a technique that analyzes the attitudes and emotions of people towards some product, service etc. Sentiment analysis of some product or service can be beneficial in predicting future scope of it. However, manually analyzing a large number of documents in a limited time can be a tedious and challenging task. Hence, several attempts have been made in the literature to solve this problem and several sentiment analysis techniques have been proposed. However, these approaches do not consider or do not give much weighted to‘emoticons’ present in the sentence. Emotions are very popular these days and have become an integral part of written communication. Hence, in this paper, we propose a novel algorithm, based on ‘emoticon score learning’ for identifying sentiment of a given sentence. We test the proposed algorithm on 1000 tweets. Experimental results show that the proposed algorithm is effective in sentiment classification and give accuracy of 91.1%. Additionally, the proposed algorithm is able to detect sentences consisting of both positive and negative sentiments.