| Author | Liu, Nan-deam |
| Call Number | AIT Thesis no. CS-95-21 |
| Subject(s) | Neural networks (Computer science)
|
| Note | A thesis submitted in partial fullfillment of the requirements for the degree of Master of
Science |
| Publisher | Asian Institute of Technology |
| Abstract | The present study was carried out in order to find a suitable method for forecasting monthly deposits and loans at the Land Bank of Taiwan (LBOT). For this purpose, Back Propagation method, Box-Jenkins Nonseasonal and Seasonal models, and Multiple Regression were used. For the Back Propagation method, trial and error were used to determine the number of nodes in the input and hidden layers. For the Box-Jenkins method, the degree of differencing was determined to obtain the smallest standard deviation of the differenced series. The selected model in each case was obtained by employing the Akaike Information Criterion and Posterior Possibility Criterion. Except for the Back Propagation method, all the remaining models employed only the time series of each deposit (or loan) itself in model development. From this study, we find out that the four different approaches for forecasting the deposit and loan activities are almost the same. Finally, to be more applicable, the weighted average of the good methods was proposed. The forecast values provided by this method would be the best compromised values to be used in practice due to the fact that they are obtained from a weighted average of the good forecast values. |
| Year | 1995 |
| Type | Thesis |
| School | School of Engineering and Technology (SET) |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Computer Science (CS) |
| Chairperson(s) | Huynh, Ngoc Phien; |
| Examination Committee(s) | Do, Ba Khang;Nagarur, Nagendra N.; |
| Scholarship Donor(s) | The Government of Republic of China; |
| Degree | Thesis (M.Sc.) - Asian Institute of Technology, 1995 |