Grade of membership model: This is the general distribution free method for the clustering of the multivariate data in which only categorical variables are included. The model assumes that the individuals can exhibit characteristics of more than one cluster, and that the state of the individual can be represented by the set or group of numerical quantities, each one equivalent to one of the clusters, which measure the 'strength' or grade of the membership of the individual for the cluster. Evaluation of these quantities and the other parameters in the model is undertaken by the maximum likelihood estimation.