Klasifikasi Sayap Lebah Apis Cerana dan Apis Koschevnikovi Menggunakan Conditional Inference Tree

Abstract

The wing of a bee is one of its biometric features. The structure of the wing venation pattern is a unique identifier that distinguishes types of bees. Venation structure looks similar so that it is difficult to distinguish by human eyes. The purpose of this study is to determine the key features on the venation pattern of bees and the accuracy of the Conditional Inference Tree method in classifying Apis cerana and Apis koschevnikovi species. In this study 19 branching points on the structure of the wing venation of bees are used as objects of observation. The data contain 89 wings of A. cerana and 239 wings of A. koschevnikovi. Each point on each species has been aligned sequentially by positions. Feature extraction was done by calculating and normalizing the distance between a pair of venation branching points. The feature extraction process produces 171 features. Then the classifier is trained and tested with Conditional Inference Tree method. Key features on the wings of bees are p77 (a pair of points 5 and 16) and p27 (a pair of points 2 and 11). Feature p76 (a pair of points 5 and 15) is an important feature used in the feature extraction stage. This method achieves accuracy of 97.78%