Evaluation of the KNN and RF methodologies for human activity recognition forecasting

Predicting the identification of human activities using k closest neighbour method compared to random forest (RF) method is the major purpose of this study work. For this study, we used G-power 0.8, alpha=0.05, and a 95% credibility range to predict the likelihood of human identification using the k-nearest neighbour (KNN) method. Our numbers of samples were 25 for Group 1 and 25 for Group 2. With an accuracy of 94.739, RF outperforms the Novel KNN. Through combining the two methods’ precision, Novel KNN outperforms RF by a wide margin.