Supervised Machine Learning Approach to COVID-19 Detection Based on Clinical Data

Abstract

Background: The new coronavirus has been spreading since the beginning of 2020, and many efforts have been made to develop vaccines to help patients recover. It is now clear that the world needs a rapid solution to curb the spread of COVID-19 worldwide with non-clinical approaches such as artificial intelligence techniques. These approaches can be effective in reducing the burden on the health care system to provide the best possible way to diagnose the COVID-19 epidemic. This study was conducted to use Machine Learning (ML) algorithms for the early detection of COVID-19 in patients. Methods: This retrospective study used data from hospitals affiliated with Shiraz University of Medical Sciences in Iran. This dataset was collected in the period March to October 2020 andcontained 10055 cases with 63 features. We selected and compared six algorithms: C4.5, support vector machine (SVM), Naive Bayes, logistic Regression (LR), Random Forest, and K-Nearest Neighbor algorithm using Rapid Miner software. The performance of algorithms was measured using evaluation metrics, such as precision, recall, accuracy, and f-measure. Results: The results of the study show that among the various used classification methods in the diagnosis of coronavirus, SVM (93.41% accuracy) and C4.5 (91.87% accuracy) achieved the highest performance. According to the C4.5 decision tree, "contact with a person who has COVID-19" was considered the most important diagnostic criterion based on the Gini index. Conclusion: We found that ML approaches enable a reasonable level of accuracy in the diagnosis of COVID-19.

Analyzing the impact of feature selection methods on machine learning algorithms for heart disease prediction

“…Artificial Intelligence (AI) technologies have advanced to a point where they offer deep, efficient, and non-intrusive analytical capabilities to facilitate the decision-making of physicians and health policy-makers in comparison to conventional methods 9 , 10 . In addition, the utilization of Machine Learning (ML) models in support of medical diagnoses, screening, and clinical prognosis is on the rise due to their high capacity to identify and categorize patients 51 .…”

“…Recent research has highlighted the potential of machine learning to improve accuracy and diagnostic time. AI-based tools constructed with machine learning have become increasingly effective diagnostic tools in recent years 9 , 10 . Machine learning algorithms are highly effective in predicting the outcome of the data in a large amount.…”

Wearable Devices and Explainable Unsupervised Learning for COVID-19 Detection and Monitoring

“…This study shows that these classical methods typically involve complex data collection, analysis, and processing, which could contribute to the overall cost and time required to perform the diagnosis process [31]. On the other hand, a supervised machine learning method can potentially induce unreliable results in COVID-19 detection and diagnosis [32] due to mislabeling of or incomplete data, which forces the model to learn incorrect patterns. Furthermore, supervised models may not generalize well to new or evolving variants of the virus, which may reduce their effectiveness in real-world scenarios.…”

Predicting Readmission of Cardiovascular Patients Admitted to the CCU using Data Mining Techniques

“…The readmission of the CV patients admitted to the CCUs of the studied hospitals within 30 days after discharge was odellin using the following algorithms through the use of the IBM SPSS Modeler 18.0 software: C5.0 [52], logistic regression [53], Decision list [54], Discriminant [55], Bayesian Network [53], K-Nearest Neighbors (KNN) [56], Random Forest [53], Support Vector Machine (SVM) [56], Chi-square Automatic Interaction Detection (CHAID) [57] The algorithms with an accuracy of >70%, namely SVM, CHIAD, artificial neural network, C5.0, KNN, logistic regression, C&R tree, and Quest algorithms, were selected.…”