On Deriving Optimal Stock-Trading Strategies in Taiwan Stock Exchange Market Based on Markov Decision Process
Autor: | Chia-Ho Chang, 張嘉合 |
---|---|
Rok vydání: | 2007 |
Druh dokumentu: | 學位論文 ; thesis |
Popis: | 95 In this study we derive optimal stock-trading strategies in Taiwan stock exchange market (TSEM) based on Markov Decision Process (MDP). First, we defined the 14 states variables in the MDP that match with the 7%-limit rule on the stock-price change in TSEM. According to historical price data at the end of each trading day, we set up the transition probability matrix for a particular stock. For each state, we defined two decisions in the MDP model, namely, either buy-in or sell-out only one unit (1,000 shares) of the stock. We calculated the expected reward for the two decisions in each state using the historical price data and obtained the optimal decision in each state by solving a linear program corresponding to the MDP model. Also, we proposed two stock-trading rules that modified the optimal strategy from the MDP model by incorporating with the rules of threshold conditions and state-variable sequence patterns. We verified the effectiveness of our optimal strategies by 110 stocks in TSEM using the real data from January of 2005 to March of 2006. We demonstrated that our optimal trading strategies out-perform other commonly-used trading strategies. |
Databáze: | Networked Digital Library of Theses & Dissertations |
Externí odkaz: |