Grade of Membership Statistical Approach
Two assumptions:
linearity:
probability that object i gives response l in the variable j is given by
conditional independence:
conditional on gik for object i responses in different variables are independent
where Yijl - indicator function that object i gives response l in the variable j
