A Framework for Detecting and Analyzing Behavior Changes of Elderly People over Time Using Learning Techniques

Area of Science:

• Gerontology
• Artificial Intelligence
• Healthcare Technology

Background:

• Sensor-rich environments offer potential for elder healthcare.
• Continuous monitoring is crucial for detecting subtle behavioral changes in the elderly.
• Existing methods often focus on short-term behavior snapshots, missing long-term trends.

Purpose of the Study:

• To develop a continuous, long-term behavioral analysis method for elder healthcare.
• To detect anomalies and changes in elderly behavior over extended periods.
• To provide actionable insights for caregivers regarding potential health conditions.

Main Methods:

• Formalized a normal behavior pattern for Activities of Daily Living (ADL).
• Developed a temporal similarity score to quantify activity changes.
• Implemented a fuzzy logic-based decision support system for disease detection and severity assessment.

Main Results:

• Successfully analyzed long-term behavior evolution to detect anomalies.
• Identified specific activities associated with detected behavior changes.
• The decision support system provides information on suspected diseases and their severity.

Conclusions:

• The proposed framework enables continuous monitoring and anomaly detection in elderly behavior.
• This approach aids in early identification of potential health issues.
• The integrated decision support system empowers caregivers with crucial health information.