The combination of Self-Organizing Feature Maps and support vector regression for solving the inverse ECG problem

Compared to body surface potentials (BSPs) recordings, myocardial transmembrane potentials (TMPs) can provide more detailed and complicated electrophysiological information. So the reconstruction of TMPs is regarded as a promising way for the diagnosis of cardiac diseases. This paper proposed the hybrid method of SVR with the Self-Organizing Feature Map (SOFM) technique to lessen training time and to improve the reconstruction accuracies. The model was implemented by the following processes: SOFM algorithm was adopted to cluster the training samples; and the individual SVR model for each cluster was then constructed. For each testing sample, find the cluster to which it belongs, and then use the corresponding SVR model to reconstruct the TMPs. The proposed model was tested and compared with single SVR schemes using a realistic heart-torso model. The experiment results show that the proposed SOFM-SVR is an improvement over the traditional single SVR in solving the inverse ECG problem, leading to a more accurate reconstruction of the TMPs.