Directional movement index based machine learning strategy for predicting stock trading signals

Abstrak

Intelligent stock trading systems are demand of the modern information age. This research paper proposed a directional movement index based machine learning (DMI-ML) strategy for predicting stock trading signals. Performance of the proposed strategy was evaluated in terms of annual rate of return (ARR), Sharpe ratio (SR), and percentage of profitable trades executed by the trading strategy. In addition, performance of the proposed model was evaluated against the strategies viz. traditional DMI, Buy-Hold. From the experimental results, we observed that the proposed strategy outperformed other strategies in terms of all three parameters. On average, the ARR obtained from the DMI-ML strategy was 52.58% higher than the ARR obtained from the Buy-Hold strategy. At the same time, the ARR of the proposed one was found 75.12% higher than the ARR obtained from the traditional DMI strategy. Furthermore, the Sharpe ratio for the DMI-ML strategy was positive for all stocks. On the other side, the percentage of profitable trades executed by the DMI-ML strategy soared in comparison to the percentage of such trades by the traditional DMI. This study also extended analysis of the proposed model with the various intelligent trading strategies proposed by authors in various literatures and concluded that the proposed DMI-ML strategy is the better strategy for stock trading.