Scholarlink

Stock market forecasting presents numerous challenges due to volatility, complexity, and nonlinear behaviour of financial markets. Recently, many deep learning models have been employed in stock market forecasting, outperforming traditional machine learning techniques. This work examines four artificial neural network models, namely, simple RNN, GRU, LSTM and Bi-LSTM for carrying out comparative study in stock market forecasting. We trained the networks using historical stock data from Standard and Poor’s 500 index and evaluated them based on performance metrics. Based on the experimental results, the LSTM model outperforms all competing configurations.

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

Comparison of Competing Artificial Neural Networks Models for Stock Market Forecasting

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Konstantinos Liagkouras,
  • Kostas Metaxiotis

摘要

Stock market forecasting presents numerous challenges due to volatility, complexity, and nonlinear behaviour of financial markets. Recently, many deep learning models have been employed in stock market forecasting, outperforming traditional machine learning techniques. This work examines four artificial neural network models, namely, simple RNN, GRU, LSTM and Bi-LSTM for carrying out comparative study in stock market forecasting. We trained the networks using historical stock data from Standard and Poor’s 500 index and evaluated them based on performance metrics. Based on the experimental results, the LSTM model outperforms all competing configurations.