ABSTRACT
Economic crises in the past have raised questions about the validity of the efficient market hypothesis and led to the development of models that can predict the stock price. The efficiency of the stock markets is known to decline, which means that the stock market’s performance is unpredictable. One of the developing models is a prediction based on economic components known to affect the IDX composite index and processing it by machine learning techniques. Support vector machines are known to have the ability to handle high-dimensional data and have advantages over other algorithms. To determine what economic components are used, this study begins with identifying the influence of domestic and international economic components on the future IDX composite index using Pearson’s correlation coefficient. The identification result explains that eight domestic and six international components have influences on the IDX composite index. Using these components as IDX composite index-influencing factors, this study builds and evaluates a machine learning-based stock index prediction model. This research used two error parameters: mean absolute error (MAE) and root mean square error (RMSE). The SVM model’s performance was compared to the most used machine learning algorithm: artificial neural networks (ANN) and the classic algorithm, multiple linear regression (MLR). The prediction results showed that SVM showed the best performance in predicting next day’s stock index prices (t+1).
