Identifikasi Jenis Shorea Berdasarkan Morfologi Daun Menggunakan Probabilistic Neural Network

Abstract

Indonesia is a country which has a large forest area. Dipterocarpaceae is the best timber tree of tropical rain forest. Shorea is a genus of Dipterocarpaceae family which consist of 194 species. Shorea has a considerable number of benefits. It is the most important timber source. Shorea can be used to produce varnish, paint and other chemical materials. Species diversity and morphological similarity may lead to difficulty in identifying the species of Shorea. Mistake in identifying Shorea can lead to inappropriate selection for the final usability. Identification using leaf is the first choice for plant classification compared to biology methods which use cell and molecule. This research aims to develop an identification system to identify Shorea based on morphological characteristic of Shorea. The system identifies 10 species of Shorea. The identification system being built uses Probabilistic Neural Network. The data are divided into five subsets. The five subsets are used as training data and test data. The PNN is trained using 80 leaves to kinds of Shorea. The identification using Probabilistic Neural Network produced 84% average accuracy.