Scholarlink

In this paper, we use Long Short-Term Memory (LSTM) models to model and forecast daily ending stock values and contrast this with a traditional method employing actual patterns of the data structure to create and train an Auto Regressive Integrated Moving Average (ARIMA) model. We gather every day’s market data—primarily Open, High, Low, And Close (OHLC) values—for companies that are listed on the FTSE and NASDAQ stock exchanges for a period of three years. Forecasts of the next day closing value are produced by both models for comparative purposes. Model performance is compared by averaging their individual Mean Squared Error (MSE) and Mean Absolute Error (MAE) scores to enable relative analysis. The results indicate that the LSTM model provides lower error rates compared to the ARIMA model, which suggests a more accurate result in forecasting stock values.

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

Comparative Analysis of LSTM and ARIMA Models for Stock Price Forecasting

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Satish Mandavalli

摘要

In this paper, we use Long Short-Term Memory (LSTM) models to model and forecast daily ending stock values and contrast this with a traditional method employing actual patterns of the data structure to create and train an Auto Regressive Integrated Moving Average (ARIMA) model. We gather every day’s market data—primarily Open, High, Low, And Close (OHLC) values—for companies that are listed on the FTSE and NASDAQ stock exchanges for a period of three years. Forecasts of the next day closing value are produced by both models for comparative purposes. Model performance is compared by averaging their individual Mean Squared Error (MSE) and Mean Absolute Error (MAE) scores to enable relative analysis. The results indicate that the LSTM model provides lower error rates compared to the ARIMA model, which suggests a more accurate result in forecasting stock values.