Research on Dairy Cow Mastitis Based on Conductance Method and Weighted Deep Forest

Abstract

Aiming at the difficult and expensive problem of dairy cow mastitis detection, an analysis method based on conductance method and weighted deep forest model is proposed, a new method of extracting conductance data features is added, and the deep random forest model is optimized by weighting. By comparing machine learning algorithms, using Accuracy, Recall, Precision, F1-Measure, and AUC (Area under Curve) as evaluation indicators, through case analysis, it is finally determined that the weighted deep forest performs well. AUC's somatic cell and somatic cell typing count reached 0.93 and 0.98, respectively.