Klasifikasi Fragmen Metagenome Menggunakan KNN dan PNN dengan Ekstraksi Fitur Gray Level Co-occurrence Matrix (GLCM) pada Variasi Jumlah Fragmen

Abstract

The development of knowledge in bioinformatics such metagenome analysis keeps evolving. Some related researches commonly use K-Mers method for the features extraction and SVM for the classification. This research uses gray level co-occurrence matrix (GLCM) method for the features extraction with KNN and PNN method for the classification. GLCM is a method to analyze the texture on image. On the DNA sequence data, the bases ACGT strand of DNA is considered as a texture with 4 levels color forms co-occurrence matrix with ACGT×ACGT size, then the texture analysis is conducted horizontally with an angle of 0 degrees. Based on this research result with the length of the fragment 200 bp, the accuracy using KNN and PNN method is 100% on the number of fragment of 1800, 18000, and 180000. From these results it can be concluded that the variation on the number of fragment does not affect the value of accuracy obtained. In addition, it can be concluded that GLCM feature extraction method can be prospectively implemented for classifying metagenome fragment.