Efficient Human Action Recognition using Novel Logistic Regression Compared over Linear Regression with Improved Accuracy

Abstract

The purpose of this study is to perform efficient human action recognition utilising novel logistic regression, as compared to linear regression, with improved accuracy. There are a total of 30 samples, which are split between the two categories. Each received a total of 15 different samples. The Linear Regression Algorithm is responsible for Group 1, whereas the Novel Logistic Regression Algorithm is in charge of Group 2.The accuracy of the model generated by the linear regression algorithm is (87.77%), which is higher than the accuracy generated by the novel logistic regression algorithm, which is (90.42%). The hypothesis is validated by the use of an independent sample T-test. The mean accuracy detection is +2SD, and the Significance Value is 0.970 (p>0.01). Both of these results indicate that the hypothesis is accurate. As a result, the accuracy of the Novel Logistic Regression Algorithm, which was measured at 90.42%, was discovered to be higher than that of the Linear Regression Algorithm, which was measured at 87.77%.