Cancer classification algorithm with improved emerging pattern

Cancer detection with gene expression profiles is important for cancer treatment.Emerging Patterns(EPs) can dis-cover the hidden gene expression patterns related to the cancer.And EPs based cancer classification algorithm can identifythe cancer for the sample.For shortcoming in the estimation of probability with frequency when the size of samples is smallin the EPs mining,Bayes estimation is introduced into the mining for increasing the reliability of entropy,then the ImprovedEmerging Patterns(EPIs) is extracted in cancer expressing,and a cancer classification algorithm with Improved Emerging Patterns(CCEPI) is proposed.The experiment is taken on the Leukemia dataset and the results show CCEPI can improve thecorrection in cancer detection.