Finance data mining1

• Main Author: Anh Cuong, Tran
• Other Authors: Hoi Chu Hong
• Format: Final Year Project
• Language: English
• Published: 2012
• Subjects: DRNTU::Engineering::Computer science and engineering::Data; DRNTU::Engineering::Computer science and engineering::Computing methodologies::Document and text processing
• Online Access: http://hdl.handle.net/10356/50844
• Institution: Nanyang Technological University

The author uses quantitative media data generated by large-scale natural language processing (NLP) text analysis systems to perform a comprehensive and comparative study on how a Twitter mood reflects its stock trading volumes. The resources for analyzing are user profile and tweets information which were collected from Twitter. Even the general results for a period of 1 year are not high (10.4%), however in some specific periods such as June 2011 or from 11th September to 21st October 2011, the accuracy outcome are about 34% and 40% respectively. This is the concrete evidence to show that media data is highly informative. Besides, while building on the findings, the author realized the gaps between today sentiment analysis tools and data related to social life, which consists of many slang, common misspellings and emotions. This might affect to the general result over one year period. In general, with a well sentiment analysis tools, predicting stock market from social media data (tweets) is possible. Improvement and recommendations for future works are also discussed.