Mastitis detection in Murrah buffaloes with intelligent models based upon electro-chemical and quality parameters of milk

Abstract

In this paper, several connectionist models have been described to detect mastitis in Murrah buffaloes using milk parameters, viz., pH, electrical conductivity, temperature (udder, milk and skin), milk somatic cells, milk yield and dielectric constant. A total of 600 milk samples were collected from 100 lactating Murrah buffaloes; which were analysed for Somatic Cell Counts in milk. Accordingly, animals were classified into three categories, i.e., healthy, subclinical mastitis and clinical mastitis animals. These basal values were utilised for developing connectionist models to identify healthy versus mastitis animals. Also, Multiple Linear Regression (MLR) models were developed for comparing classification accuracy of proposed connectionist models using Root Mean Square Error (RMSE) technique. The connectionist models were found to be superior (RMSE = 0.01) as compared to MLR models (RMSE = 4.08). Hence, it is deduced that connectionist approach could be used as a suitable technique for detecting mastitis in Murrah buffaloes.