Double threshold method for mastitis diagnosis based on NIR spectra of raw milk and chemometrics

A double threshold method to define the appropriate somatic cell count (SCC) levels for mastitis diagnosis based on the near-infrared (NIR) spectra of quarter foremilk of cows and soft independent modelling of class analogy (SIMCA) was proposed and evaluated. The accuracy of the double threshold method, as a new approach, was compared with the current single threshold method for measuring SCC for mastitis diagnosis based on SCC level in milk of individual cows. The accuracy of the SIMCA calibration model is expressed by specificity and sensitivity of disease diagnostics. As a result, 72.15% of specificity and 69.44% of sensitivity were obtained using the single threshold method which yielded its best result at 200,000 cells ml−1 SCC level. Improved accuracy was obtained with double threshold method which had 80.56% of specificity and 77.78% of sensitivity when the healthy and mastitic levels were set at 20,000 cells ml−1 and 200,000 cells ml−1, respectively.