Media Bias Detection Using Sentimental Analysis and Clustering Algorithms

Abstract

Biased media or the coverage of slanted news can have strong impact on perception of the public on topics reported by media. In the recent years, researchers have developed many comprehensive models for describing media bias effectively. These methods of analysis are manual and therefore cumbersome. Whereas in computer science and especially NLP has developed fast, automated and scalable methods which systematically analyse bias in the news posted by media houses. Various models which are generally used for analysing bias in media using computer science models appear to be simpler as compared to models developed by social science researchers. But these computer science models do not answer the most important questions despite being superior in technology. Most of the methods used in this respect are based on supervised learning and problem is that there is not enough data to go around. Also, most of the projects generally classify news as biased or unbiased. They do not tell towards which political party or ideology news is biased. In case of Indian political news data, it is not available at all. In this project, we will try to use the latest machine learning techniques like sentiment analysis, bias score and clustering to analyse the bias in the political news articles and try to group them on the basis of their media houses to predict which media houses are biased to which political parties and how much. For it, we will first collect political news article of India from various media houses using web crawlers and then filter them so that we can get only English news and separate them on basis of political alignments. After that we try to predict whether they are biased or unbiased and if they are biased, then we will try to predict towards which political party on basis of bias scores of the news article. On basis of many news articles, we will try to form a report on, which media house is biased to which political party and which media house provides unbiased news in India and see how media houses polarize public opinions.KeywordsClusteringSentiment analysisDBSCANPCAK-MeansVADER