Mining operational databases to predict potential donors among University Alumni

The University Alumni Association is currently active to ensure the excellence of Universities even though the students have commenced their studies at the University. Alumni are among the university´s most valued supporters, extend their never-ending support to the smooth running of the University and are very consistent and committed in all alumni activities arranged by the University to keep a good rapport. However, lately there has been a steep decline from some alumni member who has completely stopped giving any form of support to the University. Based on this persecution, the alumni authorities have come up with plans to lure the interests of the members to give their full commitments in contributing to the Alumni fund. One possible step taken towards this fund drive is by sending postcards to the members to make them aware of their responsibilities to the University. Based on the previous campaign data, it would like to identify those who are most likely to donate using several predictive models. This study was conducted to: (1) determine the best predictive model to predict donors; (2) identify the characteristics of donors based on the best predictive model. Two predictive models using SPSS Clementine 12 were used to predict donors, which are decision trees and logistic regression model.