Morphometric Analysis of Coleopteran Stored Product Pest using Clustering Techniques
Abstract
Insect pest species identification is a crucial process before starting any pest management program. The similar morphological characteristics among the vast insect species cause identification process difficult. In this paper, the Coleopteran stored product pest is our main concern. It causes severe damage to the stored product and gives a negative impact on its value. Therefore, K-means clustering and Hierarchical Agglomerative Cluster Analysis (HACA) were used in the identification of Coleopteran stored product pest species. Four morphological structure of 38 Coleopteran stored product pest species image was used to generate the measurement that been used as simulation data. In total, there are 100 datasets produce each morphological structure for each image. The result from K-Means Clustering produces 5 clusters while Hierarchical Agglomerative Cluster Analysis (HACA) produces 11. From the Average Silhouette index, HACA performs better in clustering compare to K-means clustering.