Comparison of different data mining algorithms used in animal science

The goal of this study was to compare predictive performance of some data mining algorithms in birth type of sheep. For this goal, CART, CHAID, Exhaustive CHAID, Naive Bayes and MLP algorithms were applied. Accuracy rate (%), sensitivity, specifity, kappa statistic and area under ROC criteria were calculated for finding the best algorithm in the predictive performance. The studied data were split into two main parts, training set (80%) and testing set (20%) for only MLP algorithm predictors The data included under the study were gender (male and female), farms (Mastung,Quetta, Noshkı) dam age and weight at lambing, lambing season and year, respectively. As a result, it was determined that all the algorithms used in the investigation showed superior performance.