DETECTING THE RELATIONSHIP OF CALIFORNIA MASTITIS TEST (CMT) WITH ELECTRICAL CONDUCTIVITY, COMPOSITION AND QUALITY OF THE MILK IN HOLSTEIN-FRIESIAN AND BROWN SWISS CATTLE BREEDS USING CART ANALYSIS

The aim of this study was to describe the relationship between mastitis and electrical conductivity, milk composition, milk quality in dairy cattle. California mastitis test (CMT) was used to make a diagnosis of mastitis in the cows, and considered as a binary response variable i.e. healthy and unhealthy. Conductivity, color measurement (L* and a*), milk fat, calving month and freezing point were considered as independent variables in the model. All the animals were classified with overall accuracy of 89.6 (%) or the error of 10.4 (%) in diagnosis of mastitis by using Classification and Regression Tree (CART) data mining algorithm. CART algorithm correctly classified unhealthy cows with an accuracy of 77.2%. The algorithm correctly classified healthy cows with a marvelous accuracy of 95.7% and a marvelous area value under ROC curve of 0.924, P=0.000). It was concluded that some measurements i.e. CMT, electrical conductivity, milk color values, milk composition and quality may be used to accurately detect mastitis together with the help of the CART algorithm.