Skip to Main Content
An approach to reconstruct the missing signals by pattern matching and neural networks is proposed in this paper for The Physionet Challenge 2010, “Mind the Gap”. The hypothesis used in this approach in the reconstruction of the missing signals is that the different cardiac signals originating from the same heart should exhibit the same signs of stress acting upon it. The level of stress in the different cardiac signals can and may vary. The neural network is built via pattern matching and cross-reference scoring of data set A. Reconstruction of the missing signal in data set B and C is based on its own prior signal data and using the trained neural network to determine the most likely segment for the filling the missing “gap”.