Evaluating RQA Characteristics of ECG Signal to Enhance Sudden Cardiac Death Prediction Time

The study presented a method to enhance SCD prediction using one of the genetic algorithms and to evaluate the characteristics of Recurrence quantification analysis (RQA). This study used data before and after the occurrence of previous attacks. The signal must be isolated a few minutes before the heart attack (SCD) to distinguish between unhealthy and normal people. In the next stage, the features related to recursive mapping were extracted using RQA. Then in the feature selection stage, the best features are selected using a genetic algorithm so that one can use them to distinguish between two groups of healthy people and those on the verge of heart death with great accuracy and predict SCD by revealing the increased risk. In the final stage, the MLP classifier was used with the help of a neural network to show the difference between the ECG signal of a healthy person and a person with SCD. Ultimately, the evaluation of feature changes according to signal RQA to improve SCD prediction needs to be considered. Performance evaluation for the DET method presents an acceptable predictive response. L method presents a good performance in prediction too. TT method has a lower performance than the three methods stated in this algorithm and is associated with estimation errors. Lmax, T1, T2, and Trans methods work almost the same, and all showed high errors in estimation in one case; if more accurate performance is required, DET and L-based evaluations are the best answer for implementing this algorithm