Use of Social Networks in Determining stockmarket Evolution

This article aims to use text mining methods and sentiment analysis to determine the stock market evolution of companies as well as virtual currencies such as Bitcoin. The source of the text is the social media channel Twitter and the text is composed of individual messages sent by users. Although previous papers proved with a degree of certainty that this paper hypothesis is true, as we will see bellow, the area of research was focused only on the professional environment or known opinion makers and not taking into account a high population mass. To ensure that a high level of information is maintained after the sentiment analysis process, we will use multiple algorithms based on different calculation methods and different word dictionaries. In addition, indicators such as the number of assessments, the number of replays etc. will be added to the methodology. By the end of the paper we will be able to both identify a working methodology of analyzing text for the purposes of stock market prediction and also we will touch on the limitations faced when creating it and the ways through which we can expand and improve it’s reliability. The implementation of all these methods and of the multiple dictionaries helped us in simulating human behavior and the differences of opinion, when a group wants to analyze a text. The algorithm becoming a way to balance the different “opinions” that resulted out of the sentiment analysis.