Diagnosis of Breast Cancer Based on Support Vector Machine and Random Forest Methods

A surge of breast cancer in health issues has set new challenges to clinical routine. With the help of data and image processing, artificial intelligence is possible to assist or automate doctors in diagnosis. On the basis of Wisconsin Breast Cancer Diagnosis (WDBC) database, this paper concentrated on Support Vector Machine (SVM) classifier and Random Forest (RF) to evaluate the features extracted from the tumor nucleus. Then the conclusion was made that SVM classifier has better prediction effect, with 97% AUC and accuracy. And the worst of the perimeter and concave points are two most important features among all the 30 characters. Hence only a model with 2 features can realize the idea results when the efficiency rate is high and data collection is simple. The outcomes showed that the promising factors to be applied to clinical diagnosis so as to suggest possible results, help laymen interpret pathological examination results, enhance the work efficiency of medical staff and mitigate their pressure.