Classifying the Emotions of Users Towards Software Products Using Sentiment Analysis

Abstract

Twitter empowers programming designers to track clients' responses to recently discharged frameworks. Such data, frequently expressed in the form of raw emotions, can be leveraged to enable a more educated programming discharge process. Nonetheless, naturally catching and translating multi-dimensional structures of human feelings communicated in Twitter messages is not a unimportant errand. Difficulties originate from the size of the information accessible, its naturally scanty nature, and the high level of domain specific words. Roused by these perceptions, in this paper we exhibit a preparatory report went for identifying, grouping, and translating feelings in programming clients' tweets. A dataset of 1000 tweets examined from an expansive scope of programming frameworks' Twitter sustains is utilized to lead our investigation. Our outcomes demonstrate that administered content classifiers (Naive Bayes and Support vector Machines) are more exact than broadly useful opinion investigation procedures in distinguishing general and specific feelings communicated in programming pertinent Tweets.

We apply Bayes methods to classify the emotions from the tweets and filter them as per the categorised words of emotions. Minig is used to filter and alter the data as per required order in order to archive so we also uses regular expressions to figure out the different emotions of the dataset.