Vol. 12, Issue 4 (2023)

Objectives: This study investigates the prevalence, risk factors, and most favourable predisposing conditions of Klebsiella mastitis in dairy goats of Punjab.

Methods: A total of 125 milk samples were collected from goats affected with clinical mastitis. The samples were processed for bacterial isolation and identification. Data regarding the mastitis status, age, bedding used, faecal contamination, weather condition, mixing with other animals were recorded and their association with Klebsiella mastitis was analysed through statistical methods and advanced machine learning models like Random Forest Classification, Gaussian Naïve Bayes Classification, XGBoost Classifier, and K-nearest neighbor Classifier Models.

Results: After molecular confirmation, the prevalence recorded was 9.6% (12/125). It was seen that all the four models recorded Sand bedding, cold weather, and age of 3 and >=3 as most appropriate conditions predisposing to Klebsiella mastitis even if they vary in their mean accuracy. However Random Forest Classification Model was the most accurate model for prediction with a Mean Accuracy of 68%.

Conclusion: Prediction modeling or Artificial intelligence are practical tools to predict conditions precisely where basic statistical techniques can fail to provide a complete picture.