Prediksi Data Runtun Waktu Menggunakan Jaringan Syaraf Tiruan

Abstract

This research concerns about application of artificial neural networks (ANN) for predicting time series data. By modifying perceptron’s activation function with linear function, we got linear networks. In order to predict time series data, these linear networks combine with adaptive LMS algorithm. And then we had completed this model with time delay function to accommodate past data in time series. The data that used in the test had varied in frequency and sampling time. Results of the test had shown that the networks work properly to predict the data series.