Summary: | 碩士 === 輔仁大學 === 資訊管理學系碩士在職專班 === 103 === This study focuses on strategic performance of Day Trading and
Swing Trading in the Taiwan Stock Exchange Capitalization Weighted Stock Index futures (TAIFEX). First, test the performance of day trading strategies with back-propagation neural network filter. Swing trading in single and multi-strategy trading strategy modules, filters, stop-loss and stop-profits, capital allocation, and the best of the mechanism of genetic algorithm parameters, if the performance of their trading strategies different?
The study showed that when the day trading strategy in RSI, KD and MACD use BPN predictions as filters, compared to not use the filter when the day trading strategy, overall performance improvement of 20%. Second, In the swing trading, the experimental results showed that the aggressive investor can Way Of The Turtle + Box Theory + combination contrarian strategy of Bollinger, the best net performance. The conservative investors can apply the Way Of The Turtle + Adam Theory + Box Theory + deviation rate theory contrarian strategy because of its risk management and control best. In this research, the use of GA to optimize the parameters of the backtesting mode, the experimental results show, GA parameter optimization strategy can effectively enhance the average 5 times higher than the net profit.
|