ANN Approach to ECG ClassificationProblem:DATAGoalsANN Approach to ECG ClassificationJoe Krachey12/10/2001ECE 539Problem:Classification of irregular heartbeats via MLP neural network.Application: – Pre-screening of patients for cardiologists–Monitoring/diagnosis equipmentData–Data provided by Massachusetts Institute of Technology and Beth Israel Hospital(MIT/BIH)–Feature vectors provided by Surehka Palreddy–Various data found at physionet.orgDATAInput vector of 9-RR0, RR1, RR2, RR1/RR0-Last 5 input vectors are threshold related.GoalsDetermine a ‘optimal’ MLP network for ECG classificationDetermine if any input vectors are not needed.Compare results with AR or KNN neural
View Full Document