Data-Driven Diagnosisof Cervical Cancer With Support Vector Machine-Based Approaches

Cervical cancer, as the fourth most common cause of death from cancer among women, has no symptoms in the early stage. There are few methods to diagnose cervical cancer precisely at present. Support vector machine (SVM) approach is introduced in this paper for cervical cancer diagnosis. Two improved SVM methods, support vector machine-recursive feature elimination and support vector machine-principal component analysis (SVM-PCA), are further proposed to diagnose the malignant cancer samples. The cervical cancer data are represented by 32 risk factors and 4 target variables: Hinselmann, Schiller, Cytology, and Biopsy. All four targets have been diagnosed and classified by the three SVM-based approaches, respectively. Subsequently, we make the comparison among these three methods and compare our ranking result of risk factors with the ground truth. It is shown that SVM-PCA method is superior to the others