Abstract:

Intelligent health analysis models improve the health sector, particularly in monitoring heart rate. Traditional unimodal systems struggle to satisfactorily provide a comprehensive and accurate prediction of heart rate deviation (HRD) as they learn from limited data features. This research study argues that an attention-based multimodal fusion model trained on authorized medical datasets remains the best possible solution, as it can manage HRD complexity. The study employs a simulation experiment methodology to compare the response accuracy of the unimodal to the attention-based multimodal model. A matrix comparison of six dataset features was selected to test a prediction of accurate and comprehensive HRD. The simulative experimental findings demonstrate that a gated attention-fused predictive multimodal system outperforms traditional unimodal systems, as heart rate deviation involves complex complementary signals.