Scholarlink

There is no financial market where currencies are traded that is as big and liquid as the foreign exchange (FX) market. This study compares machine learning algorithms for forecasting the price of USD/INR in the Forex market using 10 years of historical data, from January 2014 to January 2024. Linear Regression, Random Forest, Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU) are the four supervised learning models used in our investigation. Analyzing performance with accuracy measures, Mean Absolute Error (MAE), and Root Mean Squared Error (RMSE) produced interesting findings. As the most accurate model, Random Forest emerged, achieving an RMSE of 0.135, MAE of 0.095, and an accuracy rate of 99.95%, followed closely by Linear Regression with an RMSE of 0.057, MAE of 0.168, and an accuracy of 99.4%. However, LSTM and GRU models exhibit slightly lower accuracy, with LSTM recording an RMSE of 0.171, MAE of 0.130, and an accuracy of 97.1%, and GRU with an RMSE of 0.252, MAE of 0.200, and an accuracy of 93.7%.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Forex Market Prediction Using Machine Learning Algorithm

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • S. Adarsh,
  • Ramu Rajendran

摘要

There is no financial market where currencies are traded that is as big and liquid as the foreign exchange (FX) market. This study compares machine learning algorithms for forecasting the price of USD/INR in the Forex market using 10 years of historical data, from January 2014 to January 2024. Linear Regression, Random Forest, Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU) are the four supervised learning models used in our investigation. Analyzing performance with accuracy measures, Mean Absolute Error (MAE), and Root Mean Squared Error (RMSE) produced interesting findings. As the most accurate model, Random Forest emerged, achieving an RMSE of 0.135, MAE of 0.095, and an accuracy rate of 99.95%, followed closely by Linear Regression with an RMSE of 0.057, MAE of 0.168, and an accuracy of 99.4%. However, LSTM and GRU models exhibit slightly lower accuracy, with LSTM recording an RMSE of 0.171, MAE of 0.130, and an accuracy of 97.1%, and GRU with an RMSE of 0.252, MAE of 0.200, and an accuracy of 93.7%.