Predicting Sharia Stock Performance in Indonesia Stock Market Using Support Vector Machine algorithm for Imbalance Data
Classification, Financial ratios, Prediction, Stock’s performance, Support vector
The paper discusses the practical implementations of using the support vector machine (SVM) algorithm for imbalance data to predict the stock performances in the Indonesian stock market. SVM algorithm for imbalance data was used to model various financial ratios as independent variables to investigate indicators that significantly affect the stock’s performance of large market capitalization companies which were actively traded over the last three-year periods. The model selections, namely the imbalance and the balance SVM model with dummy variables representing the appropriate weights were carried out using 10-fold cross validation methods integrated with a grid search procedure for parameter optimization. The study identified and examined six financial ratios commonly used by the stock analysts without considering macro economic variables was able to classify the performances of the companies into two categories “good” or “poor” based on the prices proportion of two consecutive periods. The result suggested that the proposed method yield competitive performance in terms of prediction accuracy level as compared with its competitors.