A novel deep learning-based 1D-CNN-optimized GRU approach for heart disease prediction
Cardiac data modeling remains challenging in emerging nations across Asia and Africa. This research proposes an ensemble classification method leveraging machine learning (ML) to predict cardiac problems, providing physicians with actionable insights for personalized diagnoses and treatments. An ens...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Taylor & Francis Group
2025-01-01
|
| Series: | Automatika |
| Subjects: | |
| Online Access: | https://www.tandfonline.com/doi/10.1080/00051144.2024.2423430 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Cardiac data modeling remains challenging in emerging nations across Asia and Africa. This research proposes an ensemble classification method leveraging machine learning (ML) to predict cardiac problems, providing physicians with actionable insights for personalized diagnoses and treatments. An ensemble classification method for modelling cardiac temporal data is presented in this research. The minimax scalar transform is applied to first denoise the input data, and then the ENN-smote approach is applied to address the issue of an imbalanced dataset. Secondly, we employ a standard deep learning (DL) methodology. To identify the irregularities in the cardiac data pattern, a gated recurrent unit (GRU) classifier and a one-dimensional convolutional neural network (1D-CNN) are introduced. A typical genetic algorithm (GA) is used to optimize the suggested GRU network in order to pass over the local minima. This aids with 1D-CNN weight training. GA methodically optimizes the model’s GRU parameters. The data processed were finally used by the hybrid 1D-CNN-Optimized GRU network to predict cardiovascular illness. The suggested method attained a training accuracy of over 97% and a test accuracy of over 96% on the dataset. The proposed model’s overall accuracy is 99%. This is completely evaluated against other deep learning algorithms. |
|---|---|
| ISSN: | 0005-1144 1848-3380 |