Scholarlink

<p>Artificial Intelligence can be used to predict stock prices because dissatisfaction is a prerequisite to progress. Those individuals at that time dreamed of forecasting stock prices flawlessly, but this remained just a dream. Today, we use machine learning techniques to predict stock prices with high accuracy as a result of these people’s visions. It was found that the Recurrent Neural Network (RNN) is a useful tool for approximating dynamic systems. The Long Short-Term Memory (LSTM) is a RNN with multilayer cells and state memory. In this article, we highlight the importance of LSTM networks in predicting stock market movements. Time series data such as financial time series are modeled by analyzing temporal dependencies. The objective of this paper is to synthesize existing knowledge on LSTM applications, focusing on comparing different approaches to predictive modeling, and providing insight into various areas for further improvement in predictive modeling by utilizing the LSTM model to forecast the stock market. We compare 98 studies based on several important criteria. Its purpose is to inform future research and practical applications in financial decision-making by systematically reviewing these factors. In addition to LSTM networks, another contribution of the article is its investigation of the integration of additional techniques with them, such as sentiment analysis of news and social media data, which may provide an even more comprehensive view of market sentiment than LSTM networks alone. To summarize, this article presents a detailed comparative analysis that can be helpful to researchers and practitioners in the future, in addition to consolidating fragmented knowledge surrounding the use of LSTM in financial forecasting.</p>

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

Enhancing Stock Market Prediction with LSTM: A Review of Recent Developments and Comparative Analysis

  • 期刊论文

  • 发表日期：

    2025年9月3日

查看全文
  • Melika Shafiei Hafshejani,
  • Najme Mansouri

摘要

Artificial Intelligence can be used to predict stock prices because dissatisfaction is a prerequisite to progress. Those individuals at that time dreamed of forecasting stock prices flawlessly, but this remained just a dream. Today, we use machine learning techniques to predict stock prices with high accuracy as a result of these people’s visions. It was found that the Recurrent Neural Network (RNN) is a useful tool for approximating dynamic systems. The Long Short-Term Memory (LSTM) is a RNN with multilayer cells and state memory. In this article, we highlight the importance of LSTM networks in predicting stock market movements. Time series data such as financial time series are modeled by analyzing temporal dependencies. The objective of this paper is to synthesize existing knowledge on LSTM applications, focusing on comparing different approaches to predictive modeling, and providing insight into various areas for further improvement in predictive modeling by utilizing the LSTM model to forecast the stock market. We compare 98 studies based on several important criteria. Its purpose is to inform future research and practical applications in financial decision-making by systematically reviewing these factors. In addition to LSTM networks, another contribution of the article is its investigation of the integration of additional techniques with them, such as sentiment analysis of news and social media data, which may provide an even more comprehensive view of market sentiment than LSTM networks alone. To summarize, this article presents a detailed comparative analysis that can be helpful to researchers and practitioners in the future, in addition to consolidating fragmented knowledge surrounding the use of LSTM in financial forecasting.