CAJ | 학술논문

支持向量机可以引入特征变换将原空间的非线性问题转化为新空间的线性问题。本文在论述支持向量机模型创建的基础上，着重对核函数的选取及参数的确定进行了研究，通过实验数据表明，文中创建的组合核函数，在人体下肢动作模式识别中，有较高的识别率。
지지향량궤가이인입특정변환장원공간적비선성문제전화위신공간적선성문제。본문재논술지지향량궤모형창건적기출상，착중대핵함수적선취급삼수적학정진행료연구，통과실험수거표명，문중창건적조합핵함수，재인체하지동작모식식별중，유교고적식별솔。
Support vector machine ( SVM) can convert the nonlinear problem of the original space into the linear problem of new space by introducing feature transform. Based on discussing the model creation of SVM, this paper mainly studies the selection of kernel function and the determination of parameters. The experimental data shows that the combined kernel function created in this paper has higher recognition rate in human lower limb motion pattern recognition.