Research of a cluster strategy for Holter waveform
Zheng Gang · Journal of tianjin University of Technology · 2008
The paper presents a new cluster strategy for dynamic electrocardiogram(Holter) waveform,which is mainly faced with the huge amount of waveforms and distinct individual differences in Holter data.The strategy combines self-organized feature map(SOM) neural network,agglomerative clustering algorithm,k-means clustering algorithm together to reach the target that clusters the most similar waveforms in several kinds.The result shows that the strategy can get better clustering effect and higher clustering precision.