Scholarlink

<p>The study investigates long-term trends in sweet cherry (<i>Prunus avium</i>&#xa0;L.) area and production in Himachal Pradesh, India, using linear, nonlinear, autoregressive integrated moving average (ARIMA) and support vector regression (SVR) models based on annual data (from 1994 to 2024–2025). Cherry cultivation exhibited sustained expansion, with compound growth rates (CGRs) of 6.77% (area) and 11.34% (production). The cubic model best captured area and the power model best described production trends. Penalized regression splines provided the best fit (area: Akaike information criterion [AIC] = 49.99, production AIC = 58.00), outperforming locally estimated scatterplot smoothing (LOESS) and classical sigmoids. ARIMA (0,3,2) for area and ARIMA (2,2,2) for production satisfied model diagnostics. Both ARIMA and SVR showed mixed results, but SVR can be considered superior based on AIC and Bayesian information criterion (BIC). These findings provide a&#xa0;data-driven framework for policy formulation in sustainable cherry cultivation in Himachal Pradesh.</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

Predictive Modeling of Sweet Cherry (Prunus avium L.) Cultivation Area and Production in Himachal Pradesh

  • 期刊论文

  • 发表日期：

    2026年3月16日

查看全文
  • Vijit Gupta,
  • RK Gupta,
  • Ashu Chandel,
  • Geeta Verma,
  • Samriti Pathania,
  • Varnika Sharma,
  • Neha Mishra

摘要

The study investigates long-term trends in sweet cherry (Prunus avium L.) area and production in Himachal Pradesh, India, using linear, nonlinear, autoregressive integrated moving average (ARIMA) and support vector regression (SVR) models based on annual data (from 1994 to 2024–2025). Cherry cultivation exhibited sustained expansion, with compound growth rates (CGRs) of 6.77% (area) and 11.34% (production). The cubic model best captured area and the power model best described production trends. Penalized regression splines provided the best fit (area: Akaike information criterion [AIC] = 49.99, production AIC = 58.00), outperforming locally estimated scatterplot smoothing (LOESS) and classical sigmoids. ARIMA (0,3,2) for area and ARIMA (2,2,2) for production satisfied model diagnostics. Both ARIMA and SVR showed mixed results, but SVR can be considered superior based on AIC and Bayesian information criterion (BIC). These findings provide a data-driven framework for policy formulation in sustainable cherry cultivation in Himachal Pradesh.