用于动态心电波形聚类评价的Hausdorff策略

Hausdorff strategy used in clustering evaluation on AECG waveform

研究了动态心电信号的波形形态特点,提出了聚类质量有效性评价的Hausdorff策略,该策略通过类内相似度描述的相关数据的自然结构来计算并评价聚类结果的质量.通过采用MIT-BIH提供的标准心律失常数据库进行实验,实验结果和分析表明Hausdorff策略的评价结果与实际的数据分类状况呈线性相关,适用于动态心电波形聚类结果的有效性评价.

According to the study on the feature of ambulatory electrocardiogram waveform（AECG）,a cluster assessment strategy,Hausdorff is presented.The strategy calculates and evaluates the quality of clustering results through intra-cluster similarity descried in the natural structure of the data set.MIT/BIH arrhythmia data is used in the experiment,the result and analysis shows that the result of Hausdorff Stratgy is linear correlation with the real classification result,and the strategy is suitable for cluster validity evaluation on AECG data.