Research on the Quantitative Trading Strategy Based on Bollinger Band Strategy and Polynomial Regression Model

First, this paper made time-price line charts for gold and bitcoin and analyzed their primary trend of flat-rising. Subsequently, we established the Bollinger Bands strategy and a regression polynomial combinatorial model in Python to make the trade points precise. We import gold and bitcoin data into the model, respectively, and let it automatically perform graphing and data processing to obtain the fitting images, optimal fitting times, model coefficient list, and model intercept of the gold and bitcoin PR models. Afterward, we determined the trading conditions, stop-loss levels, and commissions and used the Bollinger Bands strategy to backtest the five-year trading to obtain the trading results of a single strategy. Based on the conservative principle of risk hedging and split-position trading, we reset the position distribution combined the Bollinger Bands strategy and the regression polynomial combination model into a comprehensive model. This paper plotted the fluctuations of gold and bitcoin in the time dimension. The comprehensive model shows the characteristics of high total return, low total transaction cost, and small maximum drawdown, which proves that the comprehensive model provides the best strategy.