Abstract

This article reports on the latest developments of our Emotional Stock Analyser application. The original application automatically analyses the text that is input by the user, using a text-to-emotion engine that identifies the emotive content. The whole text of newspaper articles about the Stock Market can be analysed to search for the words that indicate the "mood" of the market. The relevant emotive words relate to the up and down movement of share prices. The new application has an embedded web browser to provide the input of the text to be analysed. Users can browse the Internet and select online articles to analyse. An experiment that compared the emotions identified in web pages displaying descriptions of the previous day's movement of share prices to the next day’s Stock Market prices is described. A number of significant relationships have been observed between the emotion contained in articles about the Stock Market and the next day's market value. The output of the Online Emotion Stock Analyser was shown to be a more accurate predictor of the next day’s movement of the Stock Market share index than viewing the previous movement of the index.