Performance Analysis Of Linear Discriminant Analysis (Lda) With Optimization Techniques For Arrhythmia Classification From Ecg Signals

Document Type : Primary Research paper

Authors

1 Professor, Bannari Amman Institute Of Technology, Sathyamangalam Tamilnadu, India

2 QIP-Research Scholar, Bannari Amman Institute Of Technology, Sathayamangalam Tamilnadu, India

3 UG Scholar, PSG College Of Technology, Coimbatore, Tamilnadu, India

4 Lecturer/EEE Section, University Of Technology And Applied Sciences, Higher College Of Technology, Muscat, Oman

5 Professor, Selvam College Of Technology, Tamilnadu, India

Abstract

The Electrocardiogram (ECG) İs Monitoring The Heart’s Electrical Activity And
Pulse Rate. In Diagnosing Heart Diseases, The Analysis And Classification Of
Electrocardiogram (ECG) Records Has Become Especially Relevant. In Classifying ECG
Signals, Machine Learning Approaches Are Commonly Used. Here MIT-BIH Arrhythmia
ECG Data Base From Physionet Has Been Used To Classify Cardiogram Signals. Reduces
The Dimensionality Of Data By Using Linear Discriminant Analysis (LDA), Lastly, Well-
Known Optimization Techniques Such As Genetic Algorithm (GA), Genetic Programming
(GP) And Artificial Bee Colony (ABC) Are Used To Classify The Electrocardiogram
(ECG) Signal. The Experimental Result Analysis İndicates That The Accuracy Of GA, GP
And ABC Classifier İs 94.41 % (GA), 91.2 % (GP) And 90.8% (ABC). GA, GP And ABC
Classifiers Performance Metrics (Sensitivity (Se), Specificity (Sp) And Positive Predictivity
(Pp)) Also Compared. The Results İndicate That GA Significantly Better Performance On
All Data Sets Than GP And ABC İn Terms Of Accuracy.

Keywords