基于正则化理论的老人小孩高效SVM分类器的研究

Research on High-Efficiency SVM Classifier for the Elderly and Children Based on the Regularization Theory

由于SVM在分类过程中需要计算测试样本与所有支持向量之间的核函数,故实时性较差。所以采用基于正则化的集成线性SVM分类方法,既实现了快速分类,又能避免过拟合情况的发生,融合CNN深度学习算法更体现其良好性能。

Since the SVM needs to calculate the kernel function between the test sample and all the support vectors during the classification process,the real-time performance is poor.Therefore,the integrated linear SVM classification method based on regularization not only achieves fast classification,but also avoids the occurrence of overfitting.The fusion of CNN deep learning algorithm more reflects its good performance.