Sign Up to our social questions and Answers Engine to ask questions, answer people’s questions, and connect with other people.
Login to our social questions & Answers Engine to ask questions answer people’s questions & connect with other people.
Lost your password? Please enter your email address. You will receive a link and will create a new password via email.
Please briefly explain why you feel this question should be reported.
Please briefly explain why you feel this answer should be reported.
Please briefly explain why you feel this user should be reported.
I was searching for a clustering algorithm to fuzzy cluster categorical attributes and I found the k-modes algorithm
I’ve got the way it works but I’m not understanding if the membership or belonging matrix is calculated the same way as this matrix in fuzzy c-means algorithm?
in the "no" case would you mind to clarify the way it’s calculated?
thank you in advance for your help
Using the notation of the cited paper, the question is
is the W matrix for k-Mode calculated in the same fashion as for k-Mean ?
The answer is YES, there should be no differences at all, and of course all the constraints on W remain the same; particularly that its transpose be stochastic (*).
The main difference between k-Mean and k-Mode is with regards to the computation of Z, and of course with the distance function. I’d have to re-read the paper with a fresh mind, as ATM the details about Z are a bit fuzzy (pun intended) for me, for both k-Mode algorithms, the hard and the fuzzy one.
(*) Said less pedantically: for each object, the sum of its coefficients w for all k clusters should be 1 and all all these coefficients w must be positive (and hence in the [0,1] range).