Scholarlink

<p>The closing price of a stock is a crucial indicator for evaluating the stock market and aiding stock investors in making informed decisions. Accurately predicting the closing price of a stock is of utmost importance. This paper introduces the DW-SGA model, a hybrid prediction approach for accurately forecasting stock closing prices. It combines machine learning to select relevant stock data features with discrete wavelet decomposition, which can distinguish the high-frequency and low-frequency components in the feature sequence. The low-frequency part encapsulates overall stock price trends, while the high-frequency part captures timing details. These components are processed by stacked GRU units, creating a two-dimensional input that enhances the model’s ability to capture spatio-temporal characteristics and improve prediction accuracy. The model incorporates an attention mechanism to focus on crucial information within the hidden state sequence, enhancing its predictive capabilities. Utilizing single-step prediction, the model forecasts the closing price, and the results show superior accuracy and training efficiency, particularly for stocks with high volatility and uncertainty. This makes the DW-SGA model a valuable tool for forecasting the closing price of such stocks.</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

A hybrid prediction model for stock trend based on gated recurrent unit and wavelet transform

  • 期刊论文

  • 发表日期：

    2025年7月1日

查看全文
  • Zhuoxuan Li,
  • Youxin Wang,
  • Jinde Cao,
  • Chuangxia Huang

摘要

The closing price of a stock is a crucial indicator for evaluating the stock market and aiding stock investors in making informed decisions. Accurately predicting the closing price of a stock is of utmost importance. This paper introduces the DW-SGA model, a hybrid prediction approach for accurately forecasting stock closing prices. It combines machine learning to select relevant stock data features with discrete wavelet decomposition, which can distinguish the high-frequency and low-frequency components in the feature sequence. The low-frequency part encapsulates overall stock price trends, while the high-frequency part captures timing details. These components are processed by stacked GRU units, creating a two-dimensional input that enhances the model’s ability to capture spatio-temporal characteristics and improve prediction accuracy. The model incorporates an attention mechanism to focus on crucial information within the hidden state sequence, enhancing its predictive capabilities. Utilizing single-step prediction, the model forecasts the closing price, and the results show superior accuracy and training efficiency, particularly for stocks with high volatility and uncertainty. This makes the DW-SGA model a valuable tool for forecasting the closing price of such stocks.