Stock trading point prediction using Technical Indicators and Weighted Fuzzy Time Series

• Main Authors: Jui-hua Hsu, 許睿華
• Other Authors: Yhug-ho Leu
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/z59mu7

碩士 === 國立臺灣科技大學 === 資訊管理系 === 100 === Investments in a financial market may incur risk. According to the Capital Asset Pricing Model (CAPM), a stock with higher return rate usually exhibits higher risk. The risk of a stock is expressed in terms of the volatility of the stock price. Traditionally, the investors often use the technical analysis to determine the time points for stock transactions. This study is mainly based on the technical indicators and the weighted fuzzy time series. Borrowing the idea from the ensemble algorithms, we calculate the accuracy rate of the historical predictions of each technical indicator and the weighted fuzzy time series. With the accuracy rate as the weight, we then predict the decision as a buy, a sale or no action according to the weighted sum of the individual technical indicators. According to the experiment results, most of the predicted buy signals occur at the times when the stock prices are relatively low, while the predicted sale signals occur at the times when the stock prices are relatively high. To utilize the trading signals, we adopt the policy to continually buy a stock and to sale the stock only when the opening price of the stock is higher than the average holding price of the stock. The experiment results show that our method can acquire significant return rates from the stocks in the Taiwan stock market.