Scholarlink

<p>The present study was conducted to employ various nonlinear models for fitting lactation curves in Murrah buffaloes and to determine the most suitable nonlinear model for optimizing farm management practices. A dataset of 657 animals of first parity Murrah buffaloes were collected from daily milk yield registers maintained at Department of Livestock Production Management, LUVAS, Hisar, India. Test-day milk yield records (kg/day) were taken at monthly intervals starting from day 6th of lactation (TD1) to 275th day (TD10). The models were evaluated on the basis of goodness of fit criteria viz., Adjusted R², Root Mean Square Error (RMSE), Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC). Wood’s model showed continuous growth to peak yield and slow decline, while Brody’s exhibited a steep post-peak decline. On the basis of goodness of fitness criteria, Wood’s model, Dijkstra’s model, and the Ali and Schaeffer’s model were the most effective in describing lactation curves in Murrah buffaloes, with highest adjusted R² (0.99, 0.99 and 0.99), lowest values of RMSE (0.21, 0.08 and 0.17), AIC (-7.74, -25.52 and − 11.89), and BIC (-29.86, -47.34 and − 33.41) as compared to other models used. The Quadratic and Parabolic Exponential models exhibited a poor fit (adjusted R<sup>2</sup> &lt; 0.70), suggesting their limited usefulness in describing lactation curves. Overall, Wood’s model was considered the most suitable due to its fewer parameters, minimal computational complexity, ease of fitting and convergence, and strong biological interpretability, making it more practical compared to the Dijkstra and Ali &amp; Schaeffer 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

Modelling lactation curve in Murrah buffaloes using non-linear models

  • 期刊论文

  • 发表日期：

    2025年9月22日

查看全文
  • Rakesh Nehra,
  • Yogesh C. Bangar,
  • Kamaldeep Dhundwal,
  • Narender Singh,
  • Poonam Sevda,
  • Poonam Ratwan

摘要

The present study was conducted to employ various nonlinear models for fitting lactation curves in Murrah buffaloes and to determine the most suitable nonlinear model for optimizing farm management practices. A dataset of 657 animals of first parity Murrah buffaloes were collected from daily milk yield registers maintained at Department of Livestock Production Management, LUVAS, Hisar, India. Test-day milk yield records (kg/day) were taken at monthly intervals starting from day 6th of lactation (TD1) to 275th day (TD10). The models were evaluated on the basis of goodness of fit criteria viz., Adjusted R², Root Mean Square Error (RMSE), Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC). Wood’s model showed continuous growth to peak yield and slow decline, while Brody’s exhibited a steep post-peak decline. On the basis of goodness of fitness criteria, Wood’s model, Dijkstra’s model, and the Ali and Schaeffer’s model were the most effective in describing lactation curves in Murrah buffaloes, with highest adjusted R² (0.99, 0.99 and 0.99), lowest values of RMSE (0.21, 0.08 and 0.17), AIC (-7.74, -25.52 and − 11.89), and BIC (-29.86, -47.34 and − 33.41) as compared to other models used. The Quadratic and Parabolic Exponential models exhibited a poor fit (adjusted R2 < 0.70), suggesting their limited usefulness in describing lactation curves. Overall, Wood’s model was considered the most suitable due to its fewer parameters, minimal computational complexity, ease of fitting and convergence, and strong biological interpretability, making it more practical compared to the Dijkstra and Ali & Schaeffer models.