What is it about?
This work applies Ensemble Learning to optimization techniques. It presents a complete framework of ECG-based identification from PQRST points detection, fiducial feature calculation, feature selection, and identification. Multiple feature sets have been extracted and analyzed with high classification accuracy, and critical features have been detected.
Featured Image
Photo by Joshua Chehov on Unsplash
Why is it important?
It shows that ECG has high potential for use as next-generation biometrics with liveliness detection. Different optimization algorithms have their own strengths and weaknesses and produce different results on the same dataset. Ensemble learning can be applied to them, too, for creating a more reliable optimizer. Feature selection has been advanced by detecting multiple feature combinations with the same performance. Number of features have been significantly reduced from 71 to 24, while improving classification accuracy from 0.987 to o,9904
Perspectives
The novel use of ensemble learning at the optimization level enhances robustness by combining multiple optimization algorithms, reducing algorithmic bias, and improving generalization. Additionally, identifying multiple feature subsets with similar performance reframes feature selection as knowledge discovery, revealing that identity information is redundantly encoded across different ECG features. This research contributes to biometric security, biomedical signal processing, ensemble learning, optimization theory, feature selection, explainable AI, cybersecurity, wearable computing, and digital health systems.
Dr Mamata Pandey
Birla Institute of Technology
Read the Original
This page is a summary of: Multi-objective multi-optima ensemble binary optimization algorithm for identifying optimal set of features for ECG-based identification, Applied Soft Computing, October 2025, Elsevier,
DOI: 10.1016/j.asoc.2025.113556.
You can read the full text:
Contributors
The following have contributed to this page
