Prediction of Fleece Weight from Wool Characteristics of Sheep Using Regression Tree Method (Chaid Algorithm)

The goal of the current investigation was to predict the fleece weight from some physical wool characteristics of Akkaraman (47 heads) and Awassi ewes (108 heads) by using CHAID algorithm to obtain more flexible prediction through tree-based decision. All of the ewes were 2 years of age. The wool characteristics evaluated for the study were fleece weight (FW), staple length (SL), fiber length (FL), average number of crimps over a length of 5 cm (ANC) and wool fineness (WF). Results of the visual analysis from decision tree diagram demonstrated on the basis of CHAID algorithm that i) the Awassi sheep with both SL > 13 and FL <= 15 produced the heaviest FW average, ii) Akkaraman sheep was the group that had the lightest FW average (1.904 kg), none of all the analyzed characteristics influenced FW trait of Akkaraman sheep, iv) FL influenced solely FW of Awassi sheep with SL > 13 (Adjusted P < 0.05). To conclude, use of CHAID algorithm for the high heritable wool characteristics of economic significance could give precious information for determining sheep with high yields.