| Summary: | The purpose of this study is to compare the performance of neural network using Krzyzak algorithm and standard back propagation algorithm in forecasting domain. To implement this study a timber data set, which represents a non-seasonal time series data, is used. The performance is measured based on the accuracies, which is, quantified by root mean square error and learning speed for convergence. The results show that by using a small value of learning rate, Krzyzak algorithm is better than standard back propagation algorithm for medium and long term forecasting.
|
|---|