Detail Artificial Neural Network Modelling and Characteristics   Oleh:  Ahmad, Adang Suwandi ; Gondokaryono, Yudi S. ; Dairi, Syahirul Hakim Ad Jenis: Article from Bulletin/Magazine Dalam koleksi: Proceedings Institut Teknologi Bandung vol. 26 no. 3 (1993), page 1-14. Topik: Architecture; Neuron Models Ketersediaan Perpustakaan Pusat (Semanggi) Nomor Panggil: PP32 Non-tandon: 1 (dapat dipinjam: 0) Tandon: tidak ada     Lihat Detail Induk Isi artikel Artificial Neural Network (ANN) provides a unique computing architecture that is widely used today. ANN is a massively parallel architecture using simple adaptive elements, or neuron models, as its processing elements. Such key attributes of ANN are adaptive which means ANN can learn through experience and massively parallelism, which is a property needed to produce a high speed information processing and has potential fault tolerance properties. Presented in this paper are neuron modelling and ANN characteristic. The neuron modelling is based on the physiological examination of a biological on ANN feedfoward architecture. Two of the characteristics are the influence of the learning constant and the number of cells in the hidden layer. The capabilities to handle nonlinear plant and fault tolerance properties are also investigated. Opini Anda Klik untuk menuliskan opini Anda tentang koleksi ini! Kembali

Copyright © 2006, 2007 Unika Atma Jaya, all rights reserved