,
Amity Institute of Information Technology, Amity University , Noida, Uttar Pradesh , India
,
Assistant Professor, Department of Computer Science with Data Analytics, Sri Ramakrishna College of Arts & Science , Coimbatore, Tamil Nadu , India
,
Assistant Professor, Department of Computer science and Engineering, SRM Institute of Science and Technology , Chennai, Tamil Nadu , India
,
Assistant Professor, Department of Computer Science and Engineering in Artificial Intelligence and Machine Learning, Sridevi Women's Engineering College , Hyderabad , India
,
Department of Medicine, Termez University of Economics and Service , Termez , Uzbekistan
,
Principal, Associate Professor of Mathematics, J.K.K Nataraja College of Arts & Science , Namakkal, Tamil Nadu , India
Associate Professor, Faculty of Computing, Guru Kashi University , Punjab , India
The speed of the electronic healthcare system, clinical information system, and biomedical sensing technologies has resulted in the creation of extremely huge high-dimensional and heterogeneous medical data. Such data have substantial potential to automatically diagnose diseases, but are difficult to use because they are feature redundant, nonlinear, these models are often scalable and result in a limited interpretability of many existing models. Traditional machine learning (ML) techniques are based on manually designed features and do not always scale to high-dimensional inputs, whereas the deep learning (DL) ones, despite their mightiness, usually demand large annotated datasets and heavy computational resources. The proposed paper aims to suggest a hybrid architecture of ML and DL based on automated medical diagnosis on high-dimensional clinical and biomedical data, where deep learning based on representation learning is used together with effective classical classifiers. It consists of preprocessing (normalization, filling in of missing values, dimensionality reduction) and deep feature embedding with a hierarchical neural network and classification with optimized ML models. The proposed hybrid framework has an accuracy of 93.7, precision of 93.2, recall of 92.8, F1- score of 93.0 and AUC-ROC of 0.96, which is 3.4 to 7.3 percentage points better than standalone ML and DL models. These findings show that the hybrid design is better in diagnostic performance with less complexity of inference and scalability. The proposed system is thus seen to provide a viable and strong solution to smart e-health applications that aid in the credible automation of medical diagnosis and decision support in intelligent healthcare settings.
This is an open access article distributed under the Creative Commons Attribution Non-Commercial License (CC BY-NC) License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
0
The statements, opinions and data contained in the journal are solely those of the individual authors and contributors and not of the publisher and the editor(s). We stay neutral with regard to jurisdictional claims in published maps and institutional affiliations.