Gold Price Prediction Using Long-Short Term Memory Algorithm Based on Web Application

Dalimunthe, Rodiatul Adawiyah
Adek, Rizal Tjut
Agusniar, Cut

International Journal of Engineering, Science and Information Technology
Vol 5, No 1 (2025)

Gold is a significant investment asset, particularly in times of economic instability. Various factors, including decisions by financial authorities, inflation, and global economic dynamics, influence the fluctuations in gold prices. Accurately predicting gold prices is valuable for investors when making investment decisions. This study aims to utilize the Long Short-Term Memory (LSTM) algorithm for predicting gold prices and develop a web-based application connected to Yahoo Finance to acquire real-time gold price data. The LSTM algorithm was chosen because it handles time series data with long-term dependencies. LSTM has an architecture that allows the model to retain relevant information over long periods and forget irrelevant data. In this study, the developed LSTM model produced a Mean Absolute Error (MAE) of 19.81, indicating that the average prediction deviates by approximately 19.81 units from the actual value. Furthermore, an average Mean Absolute Percentage Error (MAPE) of 0.83% demonstrates the high prediction accuracy. The results of this study show that LSTM is an effective method for predicting gold prices. The resulting web application allows users to access gold price projections interactively, thereby assisting investors in making more accurate and data-driven decisions with easy access. Additionally, the web application offers customizable features such as adjusting prediction parameters and visualizing results in real time.  These features not only enhance user engagement but also improve decision-making processes. This research provides a practical tool for optimizing investment strategies in a dynamic economic environment by leveraging machine learning and seamless web integration.