The Author’s Methodology of Forecasting Election Results on the Basis of Big Data

   Internet mass media provides their readers with an opportunity to express their opinions on topics (comments) being published. For this purpose each article inserted in the Internet is connected with certain means of communication. These comments are taken seriously by mass media as they contain valuable information that helps find out feelings of Internet users. Social networks generate a vast amount of unstructured data due to their use by thousands of users. Therefore, our research aims at substantiation and use of machine learning (artificial intellect) approach in order to analyze users’ feelings through comments on YouTube dealing with press-conferences given by nominated candidates. Obtained results show feasibility of the approach. The author proposes other ways of its modernization.