Abstract

Bovine mastitis is caused by a diverse range of bacteria, broadly categorised as major or minor pathogens. The objective of this research was to develop an unsupervised neural network (USNN) model for detecting major and minor bacterial pathogens present in milk, based on changes in milk parameters associated with mastitis. A database of 4852 quarter milk samples with records for milk parameters and bacteriological status was used to train and validate the USNN model. Correlations (P<0.05) were found between the infection status of a quarter and its somatic cell score (SCS, 0.86), electrical resistance index (ERI, -0.59) and protein percentage (PP, 0.33). Due to significant multicolinearity, the original variables were decorrelated using principle component analysis. Sensitivity of the model for correctly detecting major and minor infections was 80% and 89%, respectively. Specificity of the model for correctly detecting non-infected cases was 97%. The model is able to differentiate infected milk from non-infected based on milk parameters associated with mastitis. It is concluded that the USNN model can be developed and incorporated into milking machines to provide a reliable basis for mastitis control.

KJ, Hassan, S Samarasinghe, and MG Lopez-Benavides

Proceedings of the New Zealand Society of Animal Production, Volume 67, Wanaka, 215-219, 2007