Scholarlink

<p>Stock market prediction plays an important role in economic decisions and informed investment, but it remains a challenging task due to the market’s nonlinear, dynamic and uncertain nature. Existing statistical and deep learning approaches often struggled with limited generalization, overfitting issues and inadequate feature representation. By motivating these issues, a novel StockGAN +  + model is introduced by integrating generative adversarial learning and graph-based modeling for stock price prediction. Here, two different types of datasets were used, such as NASDAQ and the stock ticker dataset. At the initial stage, the input data is normalized by z-score normalization for preprocessing. High-level features are extracted from the preprocessed data using a stacked autoencoder module. Based on the collected features, stock prediction is performed by the StockGAN +  + model, which combines a gated graph convolutional network (GGCN) and a temporal convolutional autoencoder (TCAE) as a discriminator and generator. The hyperparameters are dynamically tuned using improved chaotic assisted grasshopper optimization (Imp-CGop). The proposed model obtains lower MSE values of 0.0000364 and a correlation of 0.997 at the National Stock Exchange (NSE) dataset. The proposed model has obtained better performance when compared to the state-of-the-art models.</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

Stock market price prediction with efficient generative artificial intelligence with predictor network

  • 期刊论文

  • 发表日期：

    2026年5月18日

查看全文
  • Adusumalli Balaji,
  • Siddabathuni Suresh Babu,
  • Hari Krishna Deevi,
  • Vunnava Dinesh Babu,
  • Vunnam Asha Latha,
  • Popuri Srinivasarao

摘要

Stock market prediction plays an important role in economic decisions and informed investment, but it remains a challenging task due to the market’s nonlinear, dynamic and uncertain nature. Existing statistical and deep learning approaches often struggled with limited generalization, overfitting issues and inadequate feature representation. By motivating these issues, a novel StockGAN +  + model is introduced by integrating generative adversarial learning and graph-based modeling for stock price prediction. Here, two different types of datasets were used, such as NASDAQ and the stock ticker dataset. At the initial stage, the input data is normalized by z-score normalization for preprocessing. High-level features are extracted from the preprocessed data using a stacked autoencoder module. Based on the collected features, stock prediction is performed by the StockGAN +  + model, which combines a gated graph convolutional network (GGCN) and a temporal convolutional autoencoder (TCAE) as a discriminator and generator. The hyperparameters are dynamically tuned using improved chaotic assisted grasshopper optimization (Imp-CGop). The proposed model obtains lower MSE values of 0.0000364 and a correlation of 0.997 at the National Stock Exchange (NSE) dataset. The proposed model has obtained better performance when compared to the state-of-the-art models.