COVID 19 Disease Prediction: Developing a Hyper Parameter Tuning Based Machine Learning Approach


Mr.Jit Dutta & Mr.M.P Gopinath


In the healthcare industry, machine learning is commonly used to predict deadly diseases. The aim of this study was to establish and compare the output of a conventional method with a proposed system that uses the Random Forest, K-Nearest Neighbor, Decision tree, Logistic regression, and Naive Bayes classification models to predict COVID-19 disease. The proposed method assisted in tuning the hyper parameters by applying grid search and random search approaches to the five mentioned classification algorithms. The key research subject is the performance of the COVID 19 disease prediction. The hyper parameter tuned model can be used to enhance the efficiency of prediction models.In terms of accuracy, both the conventional and proposed systems were evaluated and compared. The traditional method had accuracy ranging from 43.71 percent to 91.16 percent. The proposed hyper parameter tuning model achieved accuracy levels ranging from 73.31 percent to 94.09 percent. These tests showed that the proposed prediction method is capable of producing more reliable results in predicting COVID-19 disease than the conventional method while maintaining a feasible performance.