Document Type : Original Paper
Authors
Department of Statistics, Faculty of Mathematics and Natural Sciences, Hasanuddin University, Makassar, 90245 Indonesia Makassar, Indonesia
Abstract
The Jakarta Composite Index (JCI) is a pivotal benchmark for evaluating the performance of all stocks listed on the Indonesia Stock Exchange (IDX). Given the inherent complexity, nonlinearity, and non-stationarity of stock market data, selecting robust forecasting methods is essential. This study compares the performance of the Adaptive Neuro-Fuzzy Inference System (ANFIS) and Support Vector Machine (SVM) in forecasting JCI movements. The researcher assessed prediction accuracy using Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE). The training phase revealed that the optimal ANFIS model employed the generalized bell membership function, outperforming trapezoidal and Gaussian alternatives. Concurrently, the best SVM configuration utilized a linear kernel (cost = 10), demonstrating superior performance compared to radial basis function (RBF) and sigmoid kernels. In the testing phase, ANFIS achieved an RMSE of 39.894 and MAPE of 0.4647, while SVM recorded an RMSE of 38.728 and MAPE of 0.4516. These results underscore the superior predictive capabilities of SVM, positioning it as a reliable tool for stock market forecasting. The study’s findings provide valuable insights for investors and policymakers in navigating market uncertainties and optimizing investment strategies.
Keywords
Main Subjects
|