Title: Stock Price Prediction of PT Astra International Tbk based on Support Vector Regression Approach
Authors: Alfi Nur Nitasari, Sediono, M. Fariz Fadillah Mardianto, and Elly Pusporani
Volume: 8
Issue: 12
Pages: 95-101
Publication Date: 2024/12/28
Abstract:
The capital market plays an important role in the Indonesian economy by being an indicator of investor confidence in economic conditions. PT Astra International Tbk, as one of the largest conglomerates in Indonesia, has a significant influence on the Indonesian economy. This study applies a modern nonparametric regression approach, namely Support Vector Regression (SVR) by comparing various kernel functions to predict the weekly stock price of PT Astra International from January 2021 to November 2024 obtained from investing.com. The accuracy of the model generated from various kernel functions in Support Vector Regression (SVR) is determined from the RMSE value and the MAPE value. The results showed that the Radial Basic Fuction (RBF) kernel function is more optimal as indicated by the RMSE value of 203.0671 and MAPE of 2.56% which is minimum compared to other kernel functions, so the RBF kernel function is used in prediction using testing data. The RMSE value of 155.7308 was obtained with a MAPE of 2.33% on the testing data, which indicates that the prediction with the RBF kernel function is very accurate. This research is in line with the 8th point of Sustainable Development Goals (SDGs) to increase inclusive and sustainable economic growth which is expected to be the basis for formulating policies for various parties.