Electrocardiogram Classification using Reservoir Computing

An adapted state-of-the-art method of processing information known as Reservoir Computing is used to show its utility on the open and time consuming problem of heartbeat classification. The MIT-BIH arrhythmia database is used following the guidelines of the Association for the Advancement of Medical Instrumentation (AAMI). Our approach requires no pre-processing of the electrocardiographic signal leading to a fast algorithm and approaching a real-time classification solution. Our multi-class classification results indicate an average specificity of 92.24% with an average accuracy of 86.99% showing an improvement on previously reported results. Recall and precision show an average of 80.01% and 78.65%, respectively, what makes our approach significant for its use in a clinical context.