An Improving Accuracy in Predicting the spread of Covid over Online Social Networks based on Geographical Location using Novel Autoregressive Algorithm comparing Support Vector Clustering Algorithm

Abstract

Aim: The purpose of this research work is to heighten the efficiency percent of geographical location identification to relieve the effect of covid using device studying classifiers by evaluating novel Logistic Regression algorithm and Random Forest algorithm.
Materials and Methods: Logistic Regression algorithm with sample size = 10, G-power (value=0.8)and Random Forest algorithm with
sample size = 10 were predicted many times to evaluate the efficiency percentage. Logistic Regression is evaluated by using its weights
and configurations.
Results and Discussion: Logistic Regression algorithm has better accuracy (92%) when compared to Random Forest Algorithm accuracy(21%). The results achieved with significance value p=0.680 (p>0.05) shows that two groups are statistically
insignificant.

Conclusion: Logistic Regression algorithm performed significantly better than the Random Forest algorithm.