Useful clustering outcomes from meaningful time series clustering
AusDM '07 Proceedings of the sixth Australasian conference on Data mining and analytics - Volume 70
A fuzzy clustering neural network architecture for classification of ECG arrhythmias
Computers in Biology and Medicine
| Hi-index | 0.00 |
ECG Signal is the graphical output of the electromechanical activity of heart. Any change from the normal rhythm of heart is arrhythmia. Arrhythmia interpretation and classification is one of the key areas of research. Many methods for classification of arrhythmias exist in literature. These can be broadly classified into supervised or unsupervised. The unsupervised method includes various clustering techniques. Although clustering is unsupervised type of classification, it is an advisable technique for analysis and interpretation of long term ECG Holter records. In this paper, Ascending Hierarchical clustering and k-means clustering method have been used and the results of the output of the two have been compared. The MIT-BIH arrhythmia data base has been used for the above classification and data is classified into five arrhythmia beats type i. e. Normal (N), Premature ventricular contraction (PVC), Paced beats(P), Left Bundle Branch Block(LBBB) and Right Bundle Branch Block(RBBB). Firstly the data has been pre-clustered and different data sets are made. Then both the methods are applied for clustering. It has been observed that k-means gives more success rate than Hierarchical clustering.