mean criterion value algorithm

Dear All,

I want to ask something about sequential feature selection in this below statement

"After computing the mean criterion values for each candidate feature subset, sequentialfs chooses the candidate feature subset that minimizes the mean criterion value. This process continues until adding more features does not decrease the criterion."

Anybody know about the way to calculate mean criterion values that matlab do? I mean about the root algorithm.

Thanks Adrian Tjahjana

Answer 1

[Barry Greene]

It depends on whether you are using cross validation within sequentialfs - if you are, sequentialfs will divide by the number of samples in each test set. For example if you criteriuon is misclassification rate then function (fun in matlab example) should return the number of misclassified samples - sequentialfs will then use the number of sampes in each run of cross validation to calculate the mean criterion value.