What classification algorithm to use for small unbalanced data set?

[ajg507]

Hi I am a biomedical engineering student trying to analyze a data set. I have 160 normal patients and 14 disease patients. I am trying to classify these patients based on 6 numeric predictors and 4 categorical predictors. I would like to know what kind of algorithms are suitable for this data set or if they are any that I should definitely not try. I was also wondering if V-fold cross-validation is valid for small unbalanced data sets. Thanks for the help

[lmena]

Hi I am a biomedical engineering student trying to analyze a data set. I have 160 normal patients and 14 disease patients. I am trying to classify these patients based on 6 numeric predictors and 4 categorical predictors. I would like to know what kind of algorithms are suitable for this data set or if they are any that I should definitely not try. I was also wondering if V-fold cross-validation is valid for small unbalanced data sets. Thanks for the help

Hi, I am a Computer Sciences PHD student and recently I propused a new classification algorithm (FLAIRS 2006) to cope the class imbalance problem. My algorithm named REMED (Rule Extraction for MEdical Diagnosis) generates simple rules to solve binary classification tasks, you can contact me by lmena@inaoep.mx.

Best Regards