Technical indicators are one of many tools traders and investors use to anticipate price changes of assets. In practice, these signals are observed to possess some level of predictive power used to gain excess returns. In this study, a hybrid rule based machine-learning optimized trading system is developed and tested against recent price data from the Philippine Stock Exchange. Using a Random Forest Classifier, relative price movements are forecasted with technical indicators as inputs. The outputs of the classifier are used as the basis of the technical trading strategy.

Overall, the system has been demonstrated to outperform benchmark Buy & Hold trading strategy across publicly traded stocks and indices in the market. Optimizing decision making in trading, machine learning can be used to streamline trades and maximize profitability.