FUZZY EXPERT SYSTEMS, APPLICATIONS IN LIVESTOCK AND A SAMPLE DESIGN

Abstract

The aim of this study, to introduce fuzzy logic and fuzzy expert systems, fuzzy expert systems developed in livestock and to design an expert system. As material, the heart girth, body depth and body length data of 81 hair goats grown in Siirt province were collected. Using this data, a fuzzy expert system was designed for live body weight prediction. Matlab Programme Fuzzy Logic Toolbox was used for development. The designed system works with 29 rules. The Mamdani inference method was used as inference method and the Centroid method was chosen as the defuzzification method. It was found a correlation r = 0.95 between the live body weights predicted by developed system and the live body weights data obtained by actually weighing. The determination coefficient of the model is calculated as R2 = 0.90. As a result, it was shown that fuzzy expert systems have better results in livestock with uncertainties and incomplete data and development these kinds of systems in the future was reccommended.