Abstract
A multidimensional classification method for predicting the probability of goal attainment in designing work groups was already suggested.In short, the method is called MDCP and its role is modeling conditional probabilities for the predicting.An attribute space in the MDCP is used for the modeling.The attribute space is the multidimensional space in which the selective attributes of each entity are configured as a point.Methods for selecting the attributes and a guideline on teh use of the methods are suggested from the viewpoint of the AIC minimization.