Citation
Share
Date
Abstract
Beyond being a sport, soccer has built up communities. Fans showing interest, involvement, passion and loyalty to a particular team, something known as Fan Avidity, have strengthen the sport business market. Social Networks have made incredibly easy to identify fans’commitment and expertise. Among the corpus of sport analysis, plenty of posts with a well substantiated opinion on team’s performance and reliability are wasted. Based on graph theory, social networks can be seen as a set of interconnected users with a weighted influence on its edges. Evaluating the spread influence from fans' posts retrieved from Twitter could serve as a metric for identifying fans’ intensity, if adding sentiment classification, then it is possible to score Fan Avidity. Previous work attempts to engineer new key performance indicators or apply machine learning techniques for identifying the best existing indicators, however, there is limited research on sentiment analysis. In order to achieve the Master's Degree in Computer Science, this thesis aims to strengthen a machine learning model that applies polarity and sentiment analysis on tweets, as well as discovering factors thought to be relevant on a soccer match. The final goal is to achieve a flexible mechanism which automatizes the process of gathering data before a match, with the main objective of quantifying credit on fans' sentiment along with historical factors, while evaluating soccer prediction. The left alone sentiments' model could accomplish independence from the type of tournament, league or even sport.