Hybrid CNN and XGBoost Model Tuned by Modified Arithmetic Optimization Algorithm for COVID-19 Early Diagnostics from X-ray Images

Developing countries have had numerous obstacles in diagnosing the COVID-19 worldwide pandemic since its emergence. One of the most important ways to control the spread of this disease begins with early detection, which allows that isolation and treatment could perhaps be started. According to recent results, chest X-ray scans provide important information about the onset of the infection, and this information may be evaluated so that diagnosis and treatment can begin sooner. This is where artificial intelligence collides with skilled clinicians’ diagnostic abilities. The suggested study’s goal is to make a contribution to battling the worldwide epidemic by using a simple convolutional neural network (CNN) model to construct an automated image analysis framework for recognizing COVID-19 afflicted chest X-ray data. To improve classification accuracy, fully connected layers of simple CNN were replaced by the efficient extreme gradient boosting (XGBoost) classifier, which is used to categorize extracted features by the convolutional layers. Additionally, a hybrid version of the arithmetic optimization algorithm (AOA), which is also developed to facilitate proposed research, is used to tune XGBoost hyperparameters for COVID-19 chest X-ray images. Reported experimental data showed that this approach outperforms other state-of-the-art methods, including other cutting-edge metaheuristics algorithms, that were tested in the same framework. For validation purposes, a balanced X-ray images dataset with 12,000 observations, belonging to normal, COVID-19 and viral pneumonia classes, was used. The proposed method, where XGBoost was tuned by introduced hybrid AOA, showed superior performance, achieving a classification accuracy of approximately 99.39% and weighted average precision, recall and F1-score of 0.993889, 0.993887 and 0.993887, respectively.