AI SearchThe backpropagation algorithm and group method of data handling for forecasting
Chiang, Yung Te
The backpropagation algorithm and group method of data handling for forecasting - Bangkok : Asian Institute of Technology, 1993 - 92 leaves + 1 online resource - Thesis ; no. CS-93-25 . - Asian Institute of Technology. Thesis ; no. CS-93-25 .
A thesis submitted in partial fulfillment of the requirement for the degree of Master of Science
Thesis (M.Sc.) - Asian Institute of Technology, 1993
Comparison and utilization of the common neural network with back propagation algorithm (BP) and Group Method of Data Handling (GMDH) in the forecasting of some water resources time series are attempted in the present study. For this purpose, a fully user friendly software package of GMDH was developed to facilitate the application, some modifications of BP are made to simplify its use in practical situations, and all the steps involved in both algorithms are described carefully as well. Three applications are then given for this careful comparative study and it was found that neural networks can result in better forecasting accuracy as compared to GMDH.
GMDH algorithms
The backpropagation algorithm and group method of data handling for forecasting - Bangkok : Asian Institute of Technology, 1993 - 92 leaves + 1 online resource - Thesis ; no. CS-93-25 . - Asian Institute of Technology. Thesis ; no. CS-93-25 .
A thesis submitted in partial fulfillment of the requirement for the degree of Master of Science
Thesis (M.Sc.) - Asian Institute of Technology, 1993
Comparison and utilization of the common neural network with back propagation algorithm (BP) and Group Method of Data Handling (GMDH) in the forecasting of some water resources time series are attempted in the present study. For this purpose, a fully user friendly software package of GMDH was developed to facilitate the application, some modifications of BP are made to simplify its use in practical situations, and all the steps involved in both algorithms are described carefully as well. Three applications are then given for this careful comparative study and it was found that neural networks can result in better forecasting accuracy as compared to GMDH.
GMDH algorithms