Evaluation of the effectiveness of methods of computer classification of animals in sheep breeding

A large number of livestock complicates the task of qualitative classification of animals in sheep farming. The use of modern information technologies and methods of computer data analysis is an undoubted help in solving this problem. Among these methods we can distinguish machine learning methods that have proven to be successful in many sectors of the economy. A special feature of such methods is the need to create training sets. The index selection method is also used to classify and evaluate animals. This method allows you to evaluate an animal by its own productivity based on the statistical characteristics of the analyzed sample of animals. In this case, you don't need to create training sets. It makes sense to compare the methods of machine learning and the method of index selection in the classification of sheep stock, taking into account several economically useful features. The relevance of this comparison is determined by the need to determine the effectiveness of the methods used to classify and evaluate animals. In this study, the classification of the same group of animals was performed using three different methods: discriminant analysis, decision tree method, and index selection method. Further, the criteria and indicators for the effectiveness of animal classification were determined. Based on certain criteria and indicators, the effectiveness of the classification methods used was evaluated. The reasons for classification errors are given. The study is illustrated with diagrams and tables. The conclusions obtained in the course of the work can help breeders in improving the efficiency of breeding work using information and computer technologies.