Rule induction system for banking

Rule induction is one of the major forms of data mining; it retrieves all possible interesting patterns in the database in the form of rules. The general idea is that rules are created that show the relationship between specific events in dataset. Use of rule induction system in banking allows an efficient predictive model construction, the model being based on rules, applicable for solving the problem of bank clients' possible insolvency. For credit request classification, recognition algorithm with directed learning is proposed, which is especially effective in complicated correlations search, when there is a big features set. It is based on M. Bongard's features conjunctions enumeration method. On the basis of algorithm approbation a conclusion can be made about its efficiency in crediting problem solving. It proves the efficiency of the algorithm in complicated diagnostic tasks, where the difference between different classes patterns is not obvious.