The Detection of Breast Cancer Based on Dynamic Feature Selection with EM-Bayesian Ensemble Classifier

When solving the problem in computer assisted detection by the approach of pattern recognition, the lesion data always exhibited high-dimensional and inhomogeneous, which makes most of the traditional classifiers can not performance very well. In this paper, a novel approach based on the dynamic feature subset selection and the EM algorithm with Naive Bayesian classifier integration algorithm (DSFS+EMNB) is proposed. The experimental results demonstrated that this method significantly outperforms the other methods like SVM and other traditional classification methods in terms of average accuracy, as well as generality.