Pengenalan Image Daun Menggunakan Algoritma Back Propagation dengan Fungsi Aktivasi Sigmoid Bipolar dan Tangen Hiperbolik

Abstract

Backpropagation method (back propagation) is an excellent method that used to deal with issues such as recognition of complex patterns. This method is very popular among of the other artificial neural network. Propagation or re- broadcasting is used as a term taken from the workings of this network, where the error gradient hidden layer units derived from broadcasting back errors associated with the units of output. This is because the value of the target for the hidden layer units are not given. It can be concluded that the method used to derive the gradient while minimizing the sum of squared error of network output. In this research, will be do training and recognition of the leaf, by applying back propagation neural network method. The artificial neural network is used in order to perform the training and recognition on the leaves with the aim to compare the accuracy rate of the bipolar sigmoid activation function and hyperbolic tangent activation function in the backpropagation algorithm.