Clustering analysis of Coleopteran stored product pest based on morphometric structure

Abstract

The Coleopteran stored product pest contribute severe damage to stored product. Therefore, the identification of the insect pest is crucial step in the pest management program. However, the abundant of insect pest’s species may cause the difficulty in the identification process specially when using morphological image and molecular techniques. In this paper, the identification of the insect pest species is obtained using statistical analysis which are K-means clustering and Hierarchical Agglomerative Cluster Analysis (HACA). Based on the morphometric analysis of four morphological structure of 38 Coleopteran stored product pest species image, 100 dataset is generated. As a results, from two different clustering techniques, K-Means Clustering and Hierarchical Agglomerative Cluster Analysis (HACA) produce 5 clusters and 11 clusters, respectively. From the clustering evaluation, it is show that the HACA is the best since it produce the higher average Silhouette index.