Research on the Application of Support Vector Machine Algorithm Model with Multi-Modal Data Fusion in Breast Cancer Ultrasound Image Classification

In this paper, SVM was used for breast ultrasonography. The method includes image preprocessing, feature extraction of region of interest and anomaly diagnosis of SVM classifier. First, a model based on eigenvalues is constructed by using variable precision rough sets. The fit calculation method is constructed by integrating the constraint condition of dimensionality reduction with the constraint condition. Secondly, the genes were processed by non-replay redundancy random selection, uniform hybridization and Gaussian mutation operator. Through feature extraction of CT, PET, and other samples, the corresponding feature space is constructed. The model was reduced and identified by VPRS-RUGGA-SVM. The personalized classification algorithm based on support vector machine is constructed, and the final result is obtained by voting. Experiments show that the fusion algorithm based on VPRS-RUGGA-SVM can significantly improve the generalization ability and stability of the algorithm, and VPRS-RUGGA-SVM can solve the problem of “premature convergence” well and improve the recognition effect of the algorithm. The development of statistical learning theory will be more perfect SVM classifier with high classification performance and will further improve the ability of computer-aided diagnosis.