Scholarlink

<p>The rapid economic growth of recent years has led to a surge in stock market participation, necessitating the need for accurate stock price predictions to mitigate investment risks and maximize returns. However, the dynamic nature of stock prices and their intrinsic volatility pose significant challenges to traditional statistical and machine learning (ML) models, which often struggle with overfitting, poor robustness, and limited generalization. To address these challenges, this study introduces a novel framework: EvoBagNet, an evolutionary Bagging ensemble learning model specifically designed for robust and high-accuracy stock price prediction. EvoBagNet is a scalable and efficient ensemble framework combining an Extra tree-based model, categorical boosting (CatBoost), and Light Gradient Boosting Machine (LGBM) as part of a bagging ensemble technique to enhance predictive performance. The framework incorporates Complete Empirical Mode Decomposition (CEEMD) to decompose time series data into intrinsic mode functions (IMFs) across varying frequency spectra, allowing for a more granular analysis of temporal patterns. Hyper-parameter tuning is conducted using a fast, single-objective evolutionary algorithm designed to converge efficiently on optimal configurations for the ensemble model. The framework is evaluated on datasets from nine prominent IT sector companies, employing six rigorous evaluation metrics to comprehensively assess performance. Experimental results highlight EvoBagNet’s superior accuracy, robustness, and scalability, outperforming state-of-the-art models across diverse scenarios and datasets. EvoBagNet demonstrated exceptional prediction accuracy across all datasets, achieving performance scores of 97.0% ± 0.7, 98.3% ± 0.5, 97.3% ± 0.8, 97.4% ± 0.6, 97.0% ± 1.0, 98.6% ± 0.4, 98.8% ± 0.4, 91.7% ± 1.2, and 98.4% ± 0.3 for Tech Mahindra, Mindtree, Infosys, Wipro, TCS, Mphasis, L&amp;T Tech, HCL, and Coforge, respectively. These results highlight EvoBagNet’s potential as a powerful tool for stock price forecasting, offering significant implications for informed investment strategies and financial decision-making.</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

EvoBagNet: a decomposition-driven bagging ensemble with evolutionary hyperparameter optimization for robust stock price prediction

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年2月3日

查看全文
  • Umar Bashir,
  • Kuljeet Singh,
  • Vibhakar Mansotra,
  • Akib Mohi Ud Din Khanday,
  • Mehdi Neshat

摘要

The rapid economic growth of recent years has led to a surge in stock market participation, necessitating the need for accurate stock price predictions to mitigate investment risks and maximize returns. However, the dynamic nature of stock prices and their intrinsic volatility pose significant challenges to traditional statistical and machine learning (ML) models, which often struggle with overfitting, poor robustness, and limited generalization. To address these challenges, this study introduces a novel framework: EvoBagNet, an evolutionary Bagging ensemble learning model specifically designed for robust and high-accuracy stock price prediction. EvoBagNet is a scalable and efficient ensemble framework combining an Extra tree-based model, categorical boosting (CatBoost), and Light Gradient Boosting Machine (LGBM) as part of a bagging ensemble technique to enhance predictive performance. The framework incorporates Complete Empirical Mode Decomposition (CEEMD) to decompose time series data into intrinsic mode functions (IMFs) across varying frequency spectra, allowing for a more granular analysis of temporal patterns. Hyper-parameter tuning is conducted using a fast, single-objective evolutionary algorithm designed to converge efficiently on optimal configurations for the ensemble model. The framework is evaluated on datasets from nine prominent IT sector companies, employing six rigorous evaluation metrics to comprehensively assess performance. Experimental results highlight EvoBagNet’s superior accuracy, robustness, and scalability, outperforming state-of-the-art models across diverse scenarios and datasets. EvoBagNet demonstrated exceptional prediction accuracy across all datasets, achieving performance scores of 97.0% ± 0.7, 98.3% ± 0.5, 97.3% ± 0.8, 97.4% ± 0.6, 97.0% ± 1.0, 98.6% ± 0.4, 98.8% ± 0.4, 91.7% ± 1.2, and 98.4% ± 0.3 for Tech Mahindra, Mindtree, Infosys, Wipro, TCS, Mphasis, L&T Tech, HCL, and Coforge, respectively. These results highlight EvoBagNet’s potential as a powerful tool for stock price forecasting, offering significant implications for informed investment strategies and financial decision-making.