Expert Systems For Mortgages - Goals and Objectives

Goals and Objectives

The expert system neither refuses nor grants loans, but it:

• establishes whether all the conditions for granting a particular type of loan to a given client have been satisfied, and

• calculates the required term of repayment, according to the borrower's

• means and the security to be obtained from him.

The goal is to produce applications which are correct in 80 per cent to 90 per cent of all cases, and transfer responsibility for granting or refusing loans to the branch offices.

The expert system provides the branch with a significant amount of assistance simply by producing correct applications for a loan. In many cases the client had to choose between different types of loans, and it was planned that expert system should enable bank employees to advise clients on the type of loan which best matched their needs. This, too, has been done and as such contributes to the bank employees' training.

The main tasks of expert system for mortgages focused on:

• the speed of moving a loan through red tape, which management considered to be a very important factor;

• the reduction of the errors made in the filling form;

• the shortening of the turnaround time, which was too long with classical.

Simple expert systems constitute the first phase of a loan application for mortgage purposes. After a prototype is made, the construct should be presented to expert loan officers who, working together with the knowledge engineer(s) will refine the first model. But if there is no first try which is simple and understandable, there will not be complex real-life solutions afterwards.

Whether simple or sophisticated, an expert system for mortgages should be provided with explanation facilities that show how it reaches its decisions and hence its advice. The confidence of the loan officer in the AI construct will be increased when this is done in a convincing manner.