Sentiment Analysis in political news in news recommender systems
MetadataShow full item record
Human beings are always considering the opinions of each other, especially in domains that are related to the business. It has also other applications in social science, psychology and politics. The field of politics is the scope of this thesis that has been applied in the news recommender system. Along with providing personalized news articles to the user, bringing the sentiment of politicians and news agencies about political events might be interesting for users. As there is always bias in publishing the news articles, newsreaders try to realize them manually by themselves. This thesis provides different points of view of politicians to a specific event and the name of a news agency that has published the article. Through the natural language processing methods the sentiment of politicians quotes about a specific event in news articles is realized. This analysis is based on the politician s name that is chosen by the user. The new system is designed beside the primary system of news recommender on the same page with the aim of equal situation for being interacted by the users. Later it is evaluated if this feature is increasing their interaction with the system. According to the users clicks pattern analysis and their responses to the questionnaire they are interested in using the new system instead of the primary one. Through this case study, challenges and possibilities of more development are also specified.