<![CDATA[This study aims to analyze the effectiveness of combining stock price features, technical indicators MACD and RSI and external data from Google Trends in predicting stock prices using a Machine Learning approach. Three algorithms were employed: Support Vector Regression (SVR), Multiple Linear Regression (MLR), and Multilayer Perceptron (MLP). Stock price data were obtained from Yahoo Finance, while Google Trends data were collected using keywords relevant to the Indonesia Stock Exchange (IDX). Normalization was carried out using the Z-score approach to avoid feature bias. The results show that RSI is a stable technical indicator, while MACD demonstrated a significant performance improvement after normalization. Incorporating Google Trends had a positive impact on model accuracy, particularly for MLP. Among the three models tested, MLR exhibited the most consistent performance, while MLP showed improved accuracy after preprocessing. SVR performed moderately but was sensitive to feature variations. Overall, the combination of technical indicators, public sentiment data from Google Trends, and Machine Learning algorithms provides more accurate stock price predictions. This study is expected to serve as a reference for developing investment decision support systems in capital markets.]]>
Copyrights © 2025
