Human Activity Recognition on Smartphones using Innovative Logistic Regression and Comparing Accuracy of Extra Gradient Boost Algorithm

Abstract

This work uses innovative logistic regression and extra gradient boost to compare and enhance human activity recognition for walking and sitting.Novel logistic regression and Extra Gradient Boost are applied with distinct training and testing splits to predict human activity identification.From each group, ten sets of samples are selected, yielding a total of twenty samples. About 85% of the Gpower test (g power setup parameters: α=0.05 and power=0.85, ß=0.2) comes from a T test on an independent sample.Compared to Extra Gradient Boost (90.1850%), Innovative logistic regression (95.5680%) has higher accuracy, with a statistically significant value of p = 0.001 (p < 0.05). When compared to Extra Gradient Boost, Innovative logistic regression has higher accuracy.

Keywords

• innovative logistic regression
• extra gradient boost
• machine learning
• physical abuse
• smartphones
• multiple cameras