A wavelet-based approach for time series pattern detection and events prediction applied to telemonitoring data

Area of Science:

• Cardiovascular Health
• Signal Processing
• Biomedical Engineering

Background:

• Cardiovascular diseases pose a significant health burden.
• Accurate prediction of cardiovascular events is crucial for timely intervention.
• Existing methods for cardiovascular event prediction have limitations.

Purpose of the Study:

• To develop a predictive strategy for cardiovascular events using time series analysis.
• To introduce a novel time series similarity metric based on wavelet transform.
• To enable the prediction of future blood pressure signal values for hypertension event detection.

Main Methods:

• Utilizing wavelet transform for time series analysis.
• Developing a new similarity metric to detect predefined patterns in time series data.
• Combining wavelet schemes with state-space multi-models for signal prediction.

Main Results:

• A novel time series similarity metric was successfully developed.
• The proposed methodology demonstrated capability in detecting patterns indicative of future cardiovascular events.
• The system was applied to blood pressure signals from a telemonitoring platform (TEN-HMS) to predict hypertension events.

Conclusions:

• The developed predictive strategy shows promise for early detection of cardiovascular events.
• Wavelet transform and state-space multi-models offer a robust approach for analyzing biomedical signals.
• This method can enhance remote patient monitoring and improve cardiovascular care.