User Request Emotion Prediction Approach in a Crowdsourcing Platform

Abstract

Sentiment analysis manages to recognize and arranging opinions or conclusions communicated in the source content. The far-reaching of the World Wide Web has brought another method for communicating the assumptions of people. Web-based social networking is creating an immense measure of estimation rich information as tweets, status updates, blog entries, and so on. Sentiment analysis of this user-generated information is extremely helpful in knowing the opinion of the group. Twitter sentiment analysis is troublesome contrasted with general sentiment analysis because of the existence of slang words and incorrect spellings. It is additionally a medium with a tremendous measure of data where clients can see the opinion of different clients that are categorized into various sentiment classes and are progressively developing as a key factor in dynamic. In this paper, the Emotion Prediction approach has been established by collecting the real-time twitter data from the Twitter API, pre-processing the data and to predict the emotions using various Machine Learning algorithms. The proposed user emotion prediction approach can be used to predict the emotions of the user and the user emotion approach has been compared with the machine learning and sentiment classification algorithms. With the comparative study of the algorithm, the best accuracy can be analyzed for the user emotion analysis. The crowdsourcing system can be used for the Business systems to predict the popularity of their brands and to make changes in their brand according to the feedback received. This proposed Emotion Prediction approach is mainly focusing on working with a real-time Twitter dataset. With the evaluated accuracy measure the end-user can make a quick prediction in the usage of the crowdsourcing system.