Scholarlink

Time series analysis forecasting utilizes recorded time series data to estimate future values. It is essential across various sectors counting banking, healthcare, and energy. This study elucidates ancient and contemporary time Series forecasting techniques including Moving Averages, Exponential Smoothing, ARIMA, and Seasonal ARIMA. Conventional short-term forecasting linear models, such as ARIMA, exhibit strong performance however, time-series deep learning models, including LSTM and Transformers, are more adept at managing complex non-linear correlations in time series data. Hybrid models that combine conventional and deep learning techniques exhibit superior performance.

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

Hybrid Models Combining ARIMA and Neural Networks for Time Series Forecasting

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Divyani Sen,
  • Bharat Singh Deora

摘要

Time series analysis forecasting utilizes recorded time series data to estimate future values. It is essential across various sectors counting banking, healthcare, and energy. This study elucidates ancient and contemporary time Series forecasting techniques including Moving Averages, Exponential Smoothing, ARIMA, and Seasonal ARIMA. Conventional short-term forecasting linear models, such as ARIMA, exhibit strong performance however, time-series deep learning models, including LSTM and Transformers, are more adept at managing complex non-linear correlations in time series data. Hybrid models that combine conventional and deep learning techniques exhibit superior performance.