Emergent trend detection in diurnal activity.

Area of Science:

• Gerontology
• Computer Science
• Biomedical Engineering

Background:

• Monitoring elderly individuals' daily routines is crucial for detecting health changes.
• Traditional unsupervised pattern classification methods struggle with evolving, additive datasets.
• Identifying deviations from baseline activity patterns can indicate a need for medical intervention.

Purpose of the Study:

• To develop a novel framework for analyzing temporal trends in elder daily routines.
• To address the limitations of traditional methods in adapting to new data.
• To enable early detection of potential health issues through sensor data analysis.

Main Methods:

• Utilizing feature data extracted from passive sensors.
• Implementing a framework designed to recognize temporal trends.
• Employing unsupervised learning techniques adapted for additive datasets.

Main Results:

• The proposed framework effectively identifies significant deviations from established daily routine patterns.
• The system demonstrates adaptability to continuously updating datasets.
• Aberrant consecutive days, indicative of potential health concerns, can be flagged.

Conclusions:

• The developed framework offers a robust solution for monitoring elder routines via passive sensors.
• This approach overcomes the limitations of traditional methods for time-series anomaly detection.
• Early identification of health anomalies in elders is enhanced through this adaptive monitoring system.