| Summary: | 碩士 === 國立交通大學 === 資訊科學與工程研究所 === 107 === Bitcoin trading is a hot issue in recently years. The fluctuating range of the bitcoin price is violent. Thus, bitcoin trading is high-risk and high return market. In this study, we use technical analysis which is a commonly used method in the stock market to take bitcoin as the research object and use genetic algorithm with multiple technical indicators include Stochastic Oscillator (KD) , Relative Strength Index (RSI) , Moving Average (MA) and Volume Moving Average (VMA) to produce trading strategies by training bitcoin historical data.
The bitcoin historical data is from Binance exchange and the research interval is from 1, September, 2017 to 31,January, 31, 2019. The algorithm will optimize technical indicators value to find the suitable buy-sell boundary to produce trading strategies and use the trading strategies to simulate trading on the test month. The process will train previous month data to produce the trading strategies and simulate trading on test month. The same process will continuously shift until the last month of research interval.
The result of study show that the trading strategies can maintain positive profit and its performance is better than buy-and-hold strategies and some commonly used trading strategies on stock market.
|
|---|
