Predicting body weight through biometric measurements in growing hair sheep using data mining and machine learning...

Area of Science:

• Animal Science
• Agricultural Engineering
• Data Science

Background:

• Accurate live weight determination is vital for effective herd management and sustainable livestock production.
• Traditional weighing methods can be impractical in rural settings, necessitating alternative approaches.
• Body measurements offer a potential proxy for live weight estimation in sheep.

Purpose of the Study:

• To evaluate the efficacy of data mining and machine learning algorithms for predicting live weight in hair sheep using body measurements.
• To compare the performance of Multivariate Adaptive Regression Splines (MARS), Classification and Regression Tree (CART), and Support Vector Machine Regression (SVR) algorithms.
• To identify a reliable method for estimating live weight in hair sheep, particularly where weighing tools are unavailable.

Main Methods:

• Utilized body measurement data from 280 hair sheep (2 months to 3 years old).
• Applied machine learning algorithms: MARS, CART, and SVR, with 70% for training and 30% for testing.
• Evaluated algorithm performance using goodness-of-fit criteria such as R² and r values.

Main Results:

• Both MARS and SVR algorithms demonstrated high performance, achieving similar top results for R² and r values on both training and testing datasets.
• The SVR algorithm was particularly recommended due to its strong performance across various goodness-of-fit metrics.
• The study successfully established a relationship between body measurements and live weight in hair sheep breeds.

Conclusions:

• Support Vector Machine Regression (SVR) is a highly effective machine learning tool for predicting hair sheep live weight.
• This approach can aid in establishing breed standards and potentially enhance sheep meat quality in Mexico.
• The findings support the use of machine learning for practical livestock management in resource-limited environments.