Recognition of Human activity using Machine Learning Models

Recently, smartphones have emerged as a popular research topic and are becoming increasingly important in identifying human behavior. Detailed synopsis of numerous scholarly publications on this study discusses the recognition of human action. Artificial Intelligence (AI) models are created to identify human activities from the UCI online repository. We have utilized a range of machine classification approaches, including Random Forest, kNN, Neural Network, Logistic Regression, Stochastic Gradient Descent, and Naïve Bayes, to analyze the human behavior in the multivariate data. In addition to creating AI models, the feature selection procedure reduces the dataset's dimension. For every model, a confusion matrix was created and precision and recall values were computed. The results of the experiment showed that, despite requiring more memory and processing power, Random Forest perform better at recognizing human activity than other classifiers like Logistic regression and Gradient Boosting.