Predictive Analytics Models for Health Management in Kenyan Livestock Herdsmen: A Comparative Study

Predictive analytics models are increasingly being used to improve health management in livestock herds, particularly in resource-limited settings such as Kenya. A comparative analysis was conducted using machine learning algorithms (e.g., Random Forest) with datasets from 100 randomly selected herds over a two-year period, focusing on factors such as climate conditions, dietary practices, and veterinary interventions. Random Forest models demonstrated an accuracy rate of 85% in predicting disease outbreaks, with predictive precision varying by herd size (small herds: 72%, large herds: 90%). The Random Forest model was found to be the most effective for health management prediction among the tested models. Further research should focus on validating these findings in diverse geographical and climatic conditions, with practical application recommendations provided based on this work. Model estimation used $\hat{\theta}=argmin_{\theta}\sum_i\ell(y_i,f_\theta(x_i))+\lambda\lVert\theta\rVert_2^2$, with performance evaluated using out-of-sample error.