Analysis of Supervised Feature Selection Techniques on Animal Husbandry Dataset

Data mining techniques have become an obvious need of today’s high-dimensional animal industry data. In the last decade almost every aspect of animal related activities are being captured and stored either in local or central data repositories. Due to complex animal traits such as efficiency, growth, health, stress, behavior and adaptation, data mining is an area of challenge which can be optimally performed only with reduced number of relevant features. In this paper, a comparative analysis of various feature selection techniques based on some performance measuring parameter is presented using animal husbandry dataset. This research work finds J48 classifier to perform better in comparison to other traditional classification approaches. General Terms Application of data mining techniques on animal husbandry dataset, comparative analysis of most relevant feature subset selection techniques