KPL特征提取在心电识别中的应用研究

The Research of KPL feature extraction method for ECG recognition

本文结合核方法、主元分析(PCA)和线性判别分析(LDA)等机器学习方法,提出了一种特征提取的KPL方法。本文提出的KPL方法,能够保持数据集的非线性关系和最优分类方向。使用MIT-BIH心律失常标准数据库进行实验并利用PCA、KPCA进行特征提取比较,验证了KPL方法的有效性及优势。

This paper investigates kernel method, principal component analysis （PCA） and linear discriminant analysis （LDA） algorithms for proposed KPL features extraction algorithm. This paper proposed the KPL features extraction algorithm can keep good characteristic of nonlinear relationship of data and the optimal direction of classification. Furthermore, we have realized the emulation experiments of MIT-BIH ECG Arrhythmias data base. A full comparison of features extraction by PCA, KPCA, and KPL demonstrates that KPL is effective approach.