KOMBINASI SELF ORGANIZING MAP DAN RADIAL BASIS FUNCTION NEURAL NETWORK UNTUK PEMILIHAN MODA ANGKUTAN PENUMPANG (Studi Kasus : Rute Pontianak-Kuching)

Moda is the types of transportation are available for travel. The existence of airlines serving Pontianak-Kuching route and offer services at low prices, cause the bus passengers to switch to airplane. Government or operator needs to know what mode will be selected by the passengers so that they can oversee the competition between these modes. RBF was chosen to classify the selection mode of passanger transportation Pontianak-Kuching route because of the ability of RBF in the prediction and pattern classification. RBF is a hybrid learning paradigm which trained by unsupervised in the input layer and the supervised in the output layer. RBF algorithm requires a lot of centers that make the architecture and the number of weight become inefficient. Self Organizing Map (SOM) is used to solve the problem by determining the center and the number of neurons in the hidden layer. SOM is one method of clustering with unsupervised learning that has the ability for clustering of large and less of data. The results of this study indicate that the combination of SOM-RBF better on generalized test data when compared to RBFNN and Backpropagation algorithm