Hello, I am not too sure if anyone tried this out. but I ma fitting a mixture model via the EM algorithmn. But then I tried using CAMAN package and I seem to get different results. Has anyone else had this problem? For example, I am fitting the Inness data from the text Medical Applications of Finite Mixture Models but, my aswer from my algorithmn is: Lamda1 1.728469 Lambda2 8.892720 p1 0.580639 p2 0.419361 but CAMAN gives the following results: Lamda1 1.822729 Lambda2 9.235688 p1 0.6457857 p2 0.3542143 Even the loglikelihoods I am working out is a little different from the ones supplied by CAMAN. Thanks, Jim [[alternative HTML version deleted]]