takma, çiğdem university of ege - faculty of agriculture - department of animal science, Izmir, Turkey , işçi güneri, öznur university of muğla sıtkı koçman - faculty of science - department of statistics, Muðla, Turkey , akbaş, yavuz university of ege - faculty of agriculture - department of animal science, Izmir, Turkey

Clustering of Holstein Friesians Using K-Means Method

23217

by cluster analysis units or variables can be grouped according to similarities or differences in terms of their properties. In this study total of 4496 Holstein Friesian cows were grouped two, three and four clusters according to their herd, age at first calving, lactation length and 305 day milk yields. Nonhierarchical k-means clustering technique is used for this purpose. Related traits were found statically significant for clustering of Holsteins (P 0.01). Holsteins were divided into three clusters and these clusters were found statistically different (P 0.01). The correct classification percentage of cows was 98%. In the third cluster Holsteins which have the highest milk yield, the lowest age of first calving and 305-day lactation period were proposed for breeding programs.

147

Holstein , milk yield , cluster , K , means , classification , nonhierarchical

The Journal Of Ege University Faculty Of Agriculture

151