Komparasi Fungsi Aktivasi Relu Dan Tanh Pada Multilayer Perceptron

Neural network is a popular method used in machine research, and activation functions, especially ReLu Tanh, have very important function neural networks, to minimize the error value between output layer target class. With variations number of hidden layers, as well neurons each different layer, this study analyzes 8 models classify Titanic's Survivor dataset. The result that has better performance than Tanh function, seen from average accuracy precision which higher function. addition layers no effect on increasing classification results, it can be decrease use 3 4 layers. highest was obtained model using with 50 while 100