Classifying Covid-19 X-Rays and exploring the relationships of associated data
Image classification techniques using machine learning are proving to be very efficient in classifying medical images. In this paper we will look at a public Covid-19 Image dataset accompanied by clinical data to see what we can learn from our data. The data being used is a publicly available data set of labeled chest X-Ray images along with clinical data which includes parameters such as age, gender, diagnosis, and clinical notes. Using this data, we will answer some of the following questions. What is the performance difference of popular ML models including CNNs, SVMs and KNNs? What relationships exist in the parameters found in the clinical data? Can we successfully classify other features in our dataset? Can we use a multimodal approach to increase our model performance? We will be using AUC-ROC curve to validate our findings.