Applying text mining techniques to forecast the stock market fluctuations of large it companies with twitter data: descriptive and predictive approaches to enhance the research of stock market predictions with textual and semantic data

Project presented as partial requirement for obtaining the Master’s degree in Information Management, with a specialization in Information Systems and Technologies ManagementThis research project applies advanced text mining techniques as a method to predict stock market fluctuations by merging published tweets and daily stock market prices for a set of American Information Technology companies. This project executes a systematical approach to investigate and further analyze, by using mainly R code, two main objectives: i) which are the descriptive criteria, patterns, and variables, which are correlated with the stock fluctuation and ii) does the single usage of tweets indicate moderate signal to predict with high accuracy the stock market fluctuations. The main supposition and expected output of the research work is to deliver findings about the twitter text significance and predictability power to indicate the importance of social media content in terms of stock market fluctuations by using descriptive and predictive data mining approaches, as natural language processing, topic modelling, sentiment analysis and binary classification with neural networks