Multiclass Supervised Learning Approach for SAR-COV2 Severity and Scope Prediction: SC2SSP Framework
Purpose: Identifying high-risk areas for the virus or the potential for the technique to be applied to this infectious disease might be difficult. The existing tools being used for predicting viruses exhibit various limitations. The severe pneumonia caused by the rapidly spreading coronavirus disea...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Tehran University of Medical Sciences
2025-01-01
|
| Series: | Frontiers in Biomedical Technologies |
| Subjects: | |
| Online Access: | https://fbt.tums.ac.ir/index.php/fbt/article/view/660 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Purpose: Identifying high-risk areas for the virus or the potential for the technique to be applied to this infectious disease might be difficult. The existing tools being used for predicting viruses exhibit various limitations. The severe pneumonia caused by the rapidly spreading coronavirus disease (COVID-19) is predicted to have a significant negative impact on the healthcare sector. Accurate treatment requires an urgent need for early diagnosis, which reduces pressure on the healthcare system. Computed Tomography (CT) scan and Chest X-Ray (CXR) are some of the standard image diagnoses. Although a CT scan is the most common method for diagnosis, CXR is the most frequently utilized since it is more accessible, quicker, and less expensive.
Materials and Methods: In this manuscript, the proposed model SC2SSP is a multiclass supervised learning technique that aims to predict the scope and severity of the SAR-COV2 virus using data on confirmed cases and deaths. The model may also utilize preprocessing techniques which are Gaussian smoothing for handling imbalanced data, such as oversampling or under sampling, as well as feature extraction methods such as Local Binary Pattern to identify the most relevant input features for the prediction task. Additionally, a classifier such as XGBoost can also be used to further improve the model's performance. This makes the model more robust and accurate in predicting the scope and severity of the SAR-COV2 virus.
Results: The model utilizes the Exact Greedy Algorithm to classify the spread and impact of the virus in different regions. The performance metrics like accuracy, precision, fscore and sensitivity are analyzing the proposed method performance. The proposed SC2SSP approach attains 3.101% and 7.12% higher accuracy; 24.13% and 13.04% higher precision compared with existing methods, like the Detection of COVID-19 from Chest X-ray Images Using Convolutional Neural Networks (Resnet50), Deep learning for automated recognition of covid-19 from chest X-ray images (VGGNet), respectively.
Conclusion: The conclusion and potential future healthcare planning follow the exploration of evidence-based approaches and modalities in the scope and forecast.
|
|---|---|
| ISSN: | 2345-5837 |