Seleksi Fitur Yang Berpengaruh Menggunakan Nilai Mean Pada Klasifikasi Fragmen Metagenome

Abstract

Pekuwali (2018) has conducted research into the classification of metagenome fragments using spaced k-mers. Optimize the arrangement of features using Genetic Algorithms. Pekuwali (2018) concluded that the best arrangement of features or called chromosomes is 111111110001 with a fitness value of 85.42. Chromosome 111111110001 produces 336 features of extracting DNA fragments. This research aims to find out which features influence classi fi cation and the resulting accuracy. The method used is the Mean value. The mean value method was chosen because the data distribution is normal or close to normal. This study concludes that the influential features in the classification are features 22 to 27 with an accuracy of 78.83% and features 38 to 43 with an accuracy of 79.67%.
 

Penerapan Algoritma K-Nearest Neighbors untuk Klasifikasi Fragmen Metagenom Berdasarkan Ekstraksi Fitur K-Mers

“…Tahapan utama dalam metagenom diawali dengan DNA Sequencing terhadap sampel metagenom. Data yang diambil langsung dari lingkungan memungkinkan fragmen yang dihasilkan mengandung berbagai mikroorganisme, sehingga akan berakibat pada terjadinya kesalahan perakitan terhadap fragmen metagenome (Pekuwali, Kusuma, & Buono, 2020).…”