Scholarlink

<p>This study suggests a mixed approach to predict the optimal age of hens for peak egg production by using mixed linear regression (MLR) models and through machine learning techniques. The primary objective is to enhance the accuracy and reliability of age prediction using real-world egg production data. The dataset collected from the National Root Crop Research Institute (NRCRI) chicken farm in Umudike, Nigeria, contains weekly records of hen age (in weeks) and their corresponding egg production (in grams per day). Two existing regression equations from Pandey and Kumar (J Reliab Stat Stud 7:157–168, 2014) were used as the foundation to develop a new mixed linear regression model, while five ML models Polynomial Regression (degree = 2), Decision Tree, Random Forest, XGBoost, and K-Nearest Neighbors (KNN) were employed for comparison. The performance of models were evaluated using statistical metrics such as Mean Squared Error and the coefficient of determination (R<sup>2</sup>), along with fivefold cross-validation. The results explains that the proposed MLR provides better predictive accuracy among the regression based approaches. However, among all the models, the KNN model achieved the lowest MSE score of 0.2334, indicating 28.51&#xa0;weeks as the optimal age for maximum egg production. These findings suggest that combining domain-specific regression knowledge and ML models could enhance predictive performance and could offer practical applications for data-driven poultry management. These insights can assist farm operators in optimizing feeding schedules, flock replacement, and production planning to improve overall farm productivity.</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

Machine learning models for analysis and prediction of optimal egg production

  • 期刊论文

  • 发表日期：

    2025年8月7日

查看全文
  • Saif Ali Khan,
  • Alok Kumar Shukla,
  • Subhash K. Yadav,
  • Gajendra K. Vishwakarma

摘要

This study suggests a mixed approach to predict the optimal age of hens for peak egg production by using mixed linear regression (MLR) models and through machine learning techniques. The primary objective is to enhance the accuracy and reliability of age prediction using real-world egg production data. The dataset collected from the National Root Crop Research Institute (NRCRI) chicken farm in Umudike, Nigeria, contains weekly records of hen age (in weeks) and their corresponding egg production (in grams per day). Two existing regression equations from Pandey and Kumar (J Reliab Stat Stud 7:157–168, 2014) were used as the foundation to develop a new mixed linear regression model, while five ML models Polynomial Regression (degree = 2), Decision Tree, Random Forest, XGBoost, and K-Nearest Neighbors (KNN) were employed for comparison. The performance of models were evaluated using statistical metrics such as Mean Squared Error and the coefficient of determination (R2), along with fivefold cross-validation. The results explains that the proposed MLR provides better predictive accuracy among the regression based approaches. However, among all the models, the KNN model achieved the lowest MSE score of 0.2334, indicating 28.51 weeks as the optimal age for maximum egg production. These findings suggest that combining domain-specific regression knowledge and ML models could enhance predictive performance and could offer practical applications for data-driven poultry management. These insights can assist farm operators in optimizing feeding schedules, flock replacement, and production planning to improve overall farm productivity.