An Extensive Analysis of Machine Learning Techniques With Hyper-Parameter Tuning by Bayesian Optimized SVM Kernel for the Detection of Human Lung Disease

Abstract

The COVID-19 coronavirus epidemic started in Wuhan in 2019, had a detrimental impact on millions of people’s lives, and caused many fatalities. Different limitation choices were made globally to prevent the disease, which is still present, from spreading. To stop the spread of COVID-19, it is important to use individually a medical method and automated testing technologies aimed at quick identification. Artificial intelligence (AI) methods for diagnosing COVID-19 might effectively use chest X-ray (CXR) images from chest radiography. In this research, we developed training Support Vector machine kernel option hyperparameters of the Convolutional neural networks (CNN) optimized using the Bayesian method. Additionally, CNN pre-trained models and Support Vector Machine (SVM) classifiers were used to classify the CXR images. Six well-known deep pre-trained models are compared to the optimized CNN’s performance using the CXR image dataset: COVID-19, Normal, Pneumonia Bacterial, Pneumonia Viral, and Tuberculosis (TB) for multi-classification. In this research, CNN models such as AlexNet, ReseNet50, ReseNet101, VGG16, VGG19, and InceptionV3 with the proposed method SVM kernel utilized the dataset. The proposed approach shows that using the SVM kernel could give it a better forecasting technique. ReseNet101 achieved the best accuracy of 98.7% result, a hybrid model that included Deep CNN features with Bayesian optimization produced a high degree of classification performance within a short amount of time.

Keywords

• Convolutional neural networks
• Bayesian optimization
• machine learning model
• ReseNet101
• SVM kernel