Conference Details

Abstract :

In this research work, to identify the model that may predict and forecast death or recovery because of “Covid-19” symptoms, travel history and the age of the patient, different sorts of requisite symptoms, and details related to patients should be captured correctly, also enough data should be available to coach and mentor the models of ensemble learning for accurate results. This infectious virus can be passed from an infected person's lips and nose to tiny liquid particles through the air like coughing, sneezing, speaking, or inhalation. Ensemble Learning Methods help to improve machine learning results by mixing different models. It's very useful for prediction and forecasting. In ensemble learning, different sort of methods is available consistent with the kind of problem and use cases. These learning models, primarily concentrate on a way to use the information and algorithms to simulate in the same fashion as “humans” learn and focus on how its accuracy will be increased successively. During this research, various feature selection and extraction techniques on the Covid-19 dataset are applied to select or extract the top and best features. So, the model is trained to employ a patient's past travel history, symptoms, age, and other information to predict whether the patient will endure or die from their disease (Covid-19). In this research, various ensemble algorithms are used such as Random Forest (RF), Gradient Boosting (GB), Adda Boosting (AB), and Bagging (BAG) are used. Comparative analysis on the accuracy using normal trained data set and K Fold trained data set is performed. Ensemble Algorithms RF, GB, and AB using feature selection strategies all produced highly promising results with an accuracy of 97%, and with Feature Extraction Techniques RF and Bagging also produced promising results with an accuracy of 96%.