Klasifikasi Jenis Shorea Berdasarkan Morfologi Daun Menggunakan Jaringan Syaraf Tiruan Propagasi Balik Levenberg-Marquardt

Abstract

Shorea belongs to the Dipterocarpaceae family and is one of tropical rain forest plants which is being used in timber production. Due to the existence of its various species, Shorea’s identification remains a challenge. The aim of this research was classifying Shorea’s species using Levenberg-Marquardt Backpropagation Neural Network based on 23 morphological characteristics of its leaves. This research utilized the data from 10 species of Shorea that lived in Bogor Botanical Garden, with 10 samples each. Those species were Shorea johorensis foxwf, Shorea pinanga sp, Shorea macroptera dyer, Shorea leprosula miq, Shorea lepida blume, Shorea materialis ridley, Shorea platyclados, Shorea javanica koord & val, Shorea palembanica, dan Shorea seminis. It was found that a 100% identification accuracy can be obtained with Hidden Neuron 8 and 10.