Design method of characteristic matrix used for ECG recognition

The invention requests to protect a design method of a characteristic matrix used for ECG recognition aiming at solving the problem of decreased classification precision due to ECG feature extraction methods which may lose detail information of original signals more or less. The method comprises following steps: (1) basic matrix selection, selecting a chaotic spread spectrum matrix as the basic matrix; (2) matrix optimization, successively performing normalization and discrete cosine transformation on the matrix; (3) characteristic extraction, extracting the ECG characteristics by means of the optimized matrix and performing standardization on characteristic vectors. The characteristic matrix is optimized by means of the above method; the optimized characteristic vectors extracted by the characteristic matrix are standardized for ECG recognition and classification; the classification precision can reach 94% and the precision is steady. 本发明请求保护种用于心电识别的特征矩阵设计方法，针对解决心电特征提取方法会或多或少丢失掉原始信号的细节信息的，进而导致分类精度下降的问题。该方法包括