Project Classification in Project Portfolio Management using a Multicriteria Hierarchical Discrimination Method

In investor organizations, manager's donot have standard method to discriminate more important projects. They usually select projects based on their Private estimates. But with increasing number of projects, diversity and complexity of them, the amount of error in these estimates is increased. The main issue of this research is the categorization of candidate projects, by specific criteria, into three groups. It is very important, medium and low importance.The research method is applied to the target and in terms of data collection it is an analytical-mathematical method. Since decision making on categorization requires the absolute judgment of decision makers, and the number of classes is more than two, the MHDIS modeling technique has been used among classification models because this model is firstly based on the utility functions of decision makers and the latter is specific to multi-group classification. After modeling, we tested model with 29 other projects and finally we compared the model error with the results obtained from the UTADIS model which is similar to MHDIS. The results present a lower error rate than the UTADIS model.