Deep Generative Learning-Based 1-SVM Detectors for Unsupervised COVID-19 Infection Detection Using Blood Tests

Abstract

A sample blood test has recently become an important tool to help identify false-positive/negative rRT-PCR tests. Importantly, this is mainly because it is an inexpensive and handy option to detect potential COVID-19 patients. However, this test should be conducted by certified laboratories, expensive equipment, and trained personnel, and 3-4 hours are needed to deliver results. Furthermore, it has relatively large false-negative rates around 15-20%. Consequently, an alternative and more accessible solution, quicker and less costly, is needed. This paper introduces flexible and unsupervised data-driven approaches to detect COVID-19 infection based on blood test samples. In other words, we address the problem of COVID-19 infection detection using a Blood test as an anomaly detection problem through an unsupervised deep hybrid model. Essentially, we amalgamate the features extraction capability of the Variational Autoencoder (VAE) and the detection sensitivity of the one-class Support Vector Machine algorithm (1SVM). Two sets of routine blood tests samples from the Hospital Albert Einstein, São Paulo, Brazil, and the San Raffaele Hospital in Milan, Italy, are used to assess the performance of the investigated deep learning models.Here, missing values have been imputed based on Random Forest regressor. Compared to Generative adversarial networks (GAN), Deep Belief Network (DBN), and restricted Boltzmann machine (RBM)-based 1SVM, the traditional VAE, GAN, DBN, and RBM with softmax layer as discriminator layer, and the standalone 1SVM, the proposed VAE-based 1SVM detector offers superior discrimination performance of potential COVID-19 infections. Results also revealed that the deep learning-driven 1SVM detection approaches provide promising detection performance compared to the conventional deep learning models.

EEG-Based Mental Tasks Recognition via a Deep Learning-Driven Anomaly Detector

“…Recently, deep learning models are employed to enhance medical applications in academia and industry due to their ability of extracting pertinent features of high dimensional data [ 16 ]. They demonstrated promising performance in various applications, including COVID-19 infection detection [ 17 ], Parkinson’s disease detection [ 18 ]. Various studies have investigated deep learning techniques for EEG classification problems in recent years.…”

Weather Conditions and COVID-19 Cases: Insights from the GCC Countries

“…(2022) for predicting the confirmed and recovered COVID-19 cases in India and Brazil. Finally, in a recent study by Dairi et al. (2021a) , an unsupervised detector that integrated a variational autoencoder for feature extraction with an SVM algorithm was proposed to detect COVID-19 cases using routine blood tests.…”

Using Epidemic Modeling, Machine Learning and Control Feedback Strategy for Policy Management of COVID-19

“…The use of machine learning in pandemic situations is to improve the accuracy of prediction for both infectious and non-infectious disease screening [28]. Some of the recent machine learning approaches for pandemic detection and control utilizes supervised and unsupervised learning algorithms [29]- [31]. The main limitation is that most machine learning techniques utilize only supervised learning algorithms that use pre-acquired, labeled datasets with limited emphasis on the pandemic control and policy management.…”