<![CDATA[The development of information technology has opened up new opportunities in stock market forecasting, especially in nickel commodities, which are increasingly strategic in the global energy transition. This study uses a Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), and a Gated Recurrent Unit (GRU) to forecast the movement of the S&P GSCI Nickel stock index value. Yahoo Finance time series data for the years 2018–2024 are used in the dataset. The study's findings are used to evaluate each model's capacity to forecast changes in nickel stock prices. The RNN model is used in this study because it can work with sequential information, while LSTM works with three memory gates (input, forget, output), and GRU works with 2 gates, namely update and reset. Mean Absolute Percentage Error (MAPE) presents the results of open and closed variable forecasting errors with the lowest average for the RNN model of 2.08%, the LSTM model of 2.505%, and the GRU model of 1.505%. This study is expected to contribute to investor decision-making and the identification of the most accurate forecasting model for the nickel stock index]]>
Copyrights © 2025
