Abstract
Loan assessment is an important decision-making process of commercial banks. Conventionally, the loan approval decision is made by more senior and experienced staffs of the banks. This research applies an Artificial Neural Network (ANN) approach to model the decision-making process of experienced mortgage loan assessors. The ANN is trained based on the previous assessors' judgment and learns to mimic their assessment skills. The variables used in the ANN model are quantified using fuzzy set theory to better represent loan assessment factors. Further more, the system offers benefits of reducing decision errors in loan approvals and also offers a rational means to evaluate and monitor the consistency and correctness of mortgage loan approval operation.