Application of a new evolutionary programming/adaptive boostinghybrid to breast cancer diagnosis

摘要

A new evolutionary programming/adaptive boosting (EP/AB) neuralnetwork hybrid was investigated to measure the hybrid performanceimprovement as obtained when using an EP-only derived neural network asa baseline. By combining input variables consisting of mammographylesion descriptors and patient history data, the hybrid predictedwhether the lesion was benign or malignant, which may aid in reducingthe number of unnecessary biopsies and thus the cost of mammographyscreening of breast cancer. The EP process as well as the hybrid wasoptimized using a data set of 500 biopsy-proven cases from DukeUniversity Medical Center (USA). Results showed that the hybrid provideda 15-20% classification performance improvement as measured by the ROCAz index when compared to a non-optimized EP derived architecture