Feature extraction and classification of the histopathological image plays a significant role in prediction and diagnosis of diseases, such as breast cancer. The common issues of the features matrix are that many of features may not be relevant to their diseases. Feature selection has been proved to be an effective way to improve the result of many classification methods. In this paper, an adaptive sparse support vector is proposed, with the aim of identification features, by combining the support vector machine with the weighted L1-norm. Experimental results based on a publicly recent breast cancer histopathological image datasets show that the proposed method significantly outperforms three competitor methods in terms of overall classification accuracy and the number of selected features. Thus, the proposed method can be useful for medical image classification in the real clinical practice.
Mathematics Subject Classification: 68T05
Keywords: sparse support vector machine; lasso; Wilcoxon rank sum test; histopathology image; breast cancer; feature selection.