Deep-learning detection of human cardiovascular system deterioration from pulse-wave signals

Cardiovascular diseases (CVD) are among the leading causes of mortality worldwide, making the early detection of pathological changes critically important for timely intervention. This study presents a developed neural network model designed to detect and monitor adverse changes in the human cardiovascular system through the analysis of pulse wave signals. The research utilizes a medical dataset comprising 657 data recordings from 219 patients. The dataset covers an age range of 20–89 years and includes records on disease presence and stage. The data were filtered and processed for feature extraction. After training, the model demonstrated an accuracy of 93% on test data, confirming its potential for integration into wearable devices and remote health monitoring systems.