COMPUTATIONAL SOLUTIONS FOR THE PROBLEM OF NEGATIVE SALIENCES AND NONSYMMETRY IN INDSCAL

Carroll and Chang have derived the symmetric CANDECOMP model from the INDSCAL model, to fit symmetric matrices of approximate scalar product s in the least squares sense. Typically, the CANDECOMP algorithm is us ed to estimate the parameters. In the present paper it is shown that n egative weights may occur with CANDECOMP. This phenomenon can be suppr essed by updating the weights by the Nonnegative Least Squares Algorit hm. A potential drawback of the resulting procedure is that it may pro duce two different versions of the stimulus space matrix. To obviate t his possibility, a symmetry preserving algorithm is offered, which can be monitored to produce non-negative weights as well.