Scholarlink

<p>The paper incorporates agriculture industry linkages in a stock price prediction framework. As rural India is a large market for industrial products, the paper focuses on the impact of agricultural growth on stock prices of fast-moving consumer goods (FMCG) and two-wheelers. A novel Dynamic-PSOSVR (DPSOSVR) stock price prediction framework is proposed, which extends the SVR methodology by tuning the hyperparameters using Particle Swarm Optimization (PSO) and Lévy Flight Distribution (LFD). It is executed with well-known benchmark functions, and also compared with state-of-the-art hybrid prediction models. To measure predictive efficiency, Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE), Theil’s U, and Average Relative Variance (ARV) are used. The results indicate that DPSOSVR outperformed the other predictive frameworks. Besides overall market conditions, the significant features representing the agricultural sector that affect demand for FMCG goods and two-wheelers consist of insecticide, fungicide, pesticide, and fertilizer-producing companies, and seed companies. The proposed predictive framework will be useful for manufacturing companies for developing marketing strategies, and also important for portfolio managers and mutual funds, who need to track agricultural prospects for portfolio realignment.</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 dynamic Lévy flight distribution based algorithm in SVR for stock price prediction incorporating agriculture industry linkages

  • 期刊论文

  • 发表日期：

    2026年6月23日

查看全文
  • Tanmoy Halder,
  • Souvik Ganguly,
  • Sayan Gupta,
  • Tamal Datta Chaudhuri

摘要

The paper incorporates agriculture industry linkages in a stock price prediction framework. As rural India is a large market for industrial products, the paper focuses on the impact of agricultural growth on stock prices of fast-moving consumer goods (FMCG) and two-wheelers. A novel Dynamic-PSOSVR (DPSOSVR) stock price prediction framework is proposed, which extends the SVR methodology by tuning the hyperparameters using Particle Swarm Optimization (PSO) and Lévy Flight Distribution (LFD). It is executed with well-known benchmark functions, and also compared with state-of-the-art hybrid prediction models. To measure predictive efficiency, Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE), Theil’s U, and Average Relative Variance (ARV) are used. The results indicate that DPSOSVR outperformed the other predictive frameworks. Besides overall market conditions, the significant features representing the agricultural sector that affect demand for FMCG goods and two-wheelers consist of insecticide, fungicide, pesticide, and fertilizer-producing companies, and seed companies. The proposed predictive framework will be useful for manufacturing companies for developing marketing strategies, and also important for portfolio managers and mutual funds, who need to track agricultural prospects for portfolio realignment.