MACHINE LEARNING GOES TO THE BANK

This paper describes how the APT system has been applied to loan analy sis to generalize and refine the knowledge previously used by an exper t system, in order to increase the efficiency and the compactness of t he decision rule base. The decision to lend money to industrial compan ies is a complex and risky activity for financial institutions. They n eed much expertise to deal with the large amount of information that h as to be considered for this process, and the analysis must be careful ly done in order to avoid misjudgments that would result in severe los ses of unrecoverable credit. An expert system named SPAC had been deve loped to deal with this task without fulfilling the user's expectation s. This paper presents the drawbacks of SPAC's approach and how APT, a n integrated machine learning system, has been used to acquire and ref ine domain knowledge and general decision rules from basic description s of cases provided by SPAC. The learning methodology is detailed, and a complete example of a learning session with APT is given. The final results are then compared with those obtained with SPAC.