Surinthip Sakphoowadon. Probabilistic lexicon-based approach for stock market prediction : a case study of the stock exchange of Thailand (SET). Doctoral Degree(Information Technology (International Program)). King Mongkut's University of Technology North Bangkok. Central Library. : King Mongkut's University of Technology North Bangkok, 2018.
Probabilistic lexicon-based approach for stock market prediction : a case study of the stock exchange of Thailand (SET)
Abstract:
Stock market prediction using either financial news articles or historical stock market prices has gained much attention in data mining research area. Most existing works utilize various numerical data of relevant factors which could have influenced the stock price direction such as oil price, gold price, stock prices, stock market index values, etc. in order to predict the stock price directions. Very few research papers analyze the effect of event information appearing in financial news articles on stock price movement. In this study, a Probabilistic Lexicon Based Stock Market Prediction (PLSP) algorithm is proposed as the lexicon-based predictive approach. This approach uses the proposed Thai Financial Probabilistic Lexicon (ThaiFinLex) generated from Thai financial news and stock market closing prices. There are three major steps involved in PLSP. In the first step, relevant event terms from news articles will be extracted and assigned probabilistic values to generate ThaiFinLex. In the second step, the event terms with high efficiency chosen from ThaiFinLex were further explored by using an event term analysis technique. Finally, we predict the stock price direction by using PLSP with high efficiency event terms. Our study shows that the proposed model can predict directions of the stock price movement efficiently. The PLSP can achieve a maximum of 83.33% accuracy when used to predict stock price direction of financial industry.