An improved hybrid feature reduction for increased breast cancer diagnostic performance

Purpose Early and correct diagnosis of a disease is vital for the success of treatment. Medical diagnostic decision support system can be used to improve the accuracy of the traditional diagnosis. As such, various pattern recognition methods are studied and applied to develop medical diagnostic decision support system. In this study, the effects of dimensionality reduction techniques with probabilistic neural network (PNN) on breast cancer classification are examined. Methods A hybrid method is proposed using the independent component analysis (ICA) and the discrete wavelet transform (DWT) to reduce feature vectors of Wisconsin diagnostic breast cancer (WDBC) data set. Two independent components (ICs), and one approximation coefficient of the DWT are used as a reduced feature vector to classify breast cancer using PNN. Performance measures such as accuracy, sensitivity, specificity, Youden’s index and the receiver operating characteristics (ROC) curve are computed to indicate the advantages of the hybrid feature reduction. Results The proposed feature reduction approach using ICA and DWT improves the diagnostic capability of the PNN classifier. The hybrid feature reduction has a higher diagnostic capability than the original thirty features using PNN as a classifier. Accuracy and sensitivity are 96.31% and 98.88%, while the results of the classification using the original thirty features are 92.09% and 95.52%. Conclusions Feature reduction approach using ICA and DWT together increases the performance measures of breast cancer classification using PNN, while reducing computational complexity.