RT-PCR Result Based Covid-19 Prediction Using Voting Classification Approach

Authors:

Abstract
COVID-19 infection has become a major challenge for scientists all over the world. To reduce the growth of the rapidly de veloping pandemic conditions, the optimum and most efficient way is to detect, report, isolate, diagnose and treat this disease in the early stage. This research work is based on the machine learning methods for covid-19 prediction. This paper is based on forecasting whether a patient's COVID-19 results will be positive or negative. Each pending patient record leaves a time frame when it's unclear whether a result will be positive or negative as laboratory results are processed. The Covid-19 prediction techniques have various phases which include data input, pre-processing, feature extraction, and classification. An authentic dataset is collected for the prediction of covid 19 in this study.This research work has proposed a voting classification method, which is the combination of Gaussian naive Bayes, Bernoulli naive Bayes, random forest, and SVM. In the voting classification method data, which is taken as input, after the pre-processing PCA algorithm is applied for the feature reduction. The data with the reduced features will be given as input to k-means for the clustering. The last voting classification method is applied for the COVID-19 prediction.The proposed model is implemented in Python, and the outcomes are evaluated with respect to precision, recall, and accuracy. In comparison to previous classification techniques, the suggested model has a high level of accuracy, precision, and recall.

This publication has 17 references indexed in Scilit: