Scholarlink

The study compares the forecasting performance of two prominent time series models, ARIMA and LSTM Bidirectional, using data on crude oil prices. The evaluation metrics, including RMSE, MAE, and MAPE, indicate that the LSTM Bidirectional model consistently outperforms the ARIMA(0,1,0)(0,0,0) model across all measures (RMSE: 91.36 vs. 266.64, MAE: 67.487 vs. 225.35, MAPE: 19 vs. 43). These results underscore the LSTM Bidirectional model's superior accuracy in capturing and predicting the complex patterns inherent in crude oil price fluctuations. The findings suggest that the LSTM Bidirectional model offers a more robust and reliable forecasting solution for this specific time series dataset than the traditional ARIMA approach.

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

Forecasting Crude Oil Prices with ARIMA and LSTM Bidirectional Models—A Comparative Analysis

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • M. Nagesh,
  • D. Mallikarjuna Reddy,
  • Y. Vijayakumar,
  • M. Rajendar

摘要

The study compares the forecasting performance of two prominent time series models, ARIMA and LSTM Bidirectional, using data on crude oil prices. The evaluation metrics, including RMSE, MAE, and MAPE, indicate that the LSTM Bidirectional model consistently outperforms the ARIMA(0,1,0)(0,0,0) model across all measures (RMSE: 91.36 vs. 266.64, MAE: 67.487 vs. 225.35, MAPE: 19 vs. 43). These results underscore the LSTM Bidirectional model's superior accuracy in capturing and predicting the complex patterns inherent in crude oil price fluctuations. The findings suggest that the LSTM Bidirectional model offers a more robust and reliable forecasting solution for this specific time series dataset than the traditional ARIMA approach.