An efficient algorithm for solving a quadratic programming model with application in credit card holders' behavior

A quadratic programming model is developed for predicting the credit card holders' behaviors. This quadratic program is solved by a new algorithm. This algorithm optimizes the objective function along certain directions in the null space of the constraint equations. In this work, the results obtained from the model are compared with the results obtained from the other models for credit card holders' behavior problems. Due to solution algorithm and nature of the model the results are efficient and promising.