Publication

Abstract : Sentiment analysis is a process of identifying and extracting subjective information in source materials by performing text analysis and Natural Language Processing. It aims to determine the attitude of a speaker or a writer with respect to some topic or the overall contextual polarity of a document. Conventionally, a machine learning algorithm is applied to classify the polarity of a given text into positive, negative or neutral and this classification is done based on emotional states such as „angry‟, „sad‟, „happy‟ etc. A better classification can be achieved by considering emoticons along with emotional states. In many past studies, the emoticons played an important role in building sentiment lexicons and in training machine learning classifiers and also they are considered to be the reliable indicators of sentiment. But, real meaning of all emoticons is not known to many of the social media users. Clustering of words and emoticons in the context of social media will give a good insight about the meaning conveyed by the emoticons. Emoticons are labeled as positive, negative or neutral based on the respective cluster of words they come under. Emoticons are identified from the text and sentiment analysis is performed by using emotional states in the text and emoticons. Such an analysis will result in better classification. This paper focuses on clustering of words and emoticons to know the meaning conveyed by the emoticons and compare the results of sentiment analysis before and after the emoticons are removed from the text. © 2016, International Journal of Pharmacy and Technology. All rights reserved.