A Comparative Study of Deep Learning Architectures for Activity Recognition

This is the authors' abstract. We don't add key points for this paper.

This study introduces an advanced deep learning model, showcasing its superior performance against established architectures like CNNs, LSTMs, and Transformers in the realm of predictive analytics. The Proposed model outperforms existing benchmarks, achieving an 84.09% accuracy, 86% precision, and 85% F1-score, while substantially reducing the False Rejection Rate to 16.53%. Its robustness is validated by a consistent Receiver Operating Characteristic score of 0.99. The results advocate for the model’s application in advanced predictive tasks, setting a new standard for accuracy and reliability.