An Optimal Model Combining SqueezeNet and Machine Learning Methods for Lung Disease Diagnosis

aims: Proposing a new AI model that allows the separation, diagnosis, and classification of three diseases. background: Artificial intelligence (AI) is being increasingly applied in various fields, and it has the potential to bring about radical transformations in all areas of life, particularly in medical imaging. In the field of medicine, AI can help patients diagnose their medical conditions online before visiting a doctor, and it can also assist medical professionals in offering more accurate diagnoses and treatments. Many datasets provide a solid foundation for learning and improving machine learning methods to identify and detect certain diseases. However, some diseases, such as lung diseases detected using chest X-rays (CXR), are often very similar and challenging to detect. objective: The separation, diagnosis, and classification of three diseases. method: The proposed method is based on a combination of deep learning using the SqueezeNet architecture and machine learning algorithms : SVM, KNN, logistic regression, decision tree, and Naïve Bayes. result: The obtained results demonstrate that the proposed model is a powerful tool for the multi-class classification of lung diseases. conclusion: The proposed model provides better diagnosis and identification of diseases, with a saving of time and server space. other: The proposed model is applied to a chest X-ray dataset containing CXR images divided into four classes : Pneumonia, Tuberculosis, Covid-19, and Normal cases.