A unified grid-based wandering pattern detection algorithm.

Area of Science:

• Gerontology and Health Informatics
• Biomedical Engineering
• Data Science

Background:

• Wandering behavior is a significant concern in various clinical populations, impacting patient safety and care.
• Previous research has often focused on either indoor or outdoor wandering detection separately.
• A unified approach for detecting both indoor and outdoor wandering patterns has been lacking.

Purpose of the Study:

• To develop and validate a robust algorithm for detecting both indoor and outdoor wandering patterns.
• To analyze the relationship between identified wandering patterns and other clinical measures.
• To address the limitations of existing methods by proposing a unified algorithmic solution.

Main Methods:

• Development of a novel grid-based layout representation strategy for pattern identification.
• Application of the algorithm to both indoor and outdoor environments.
• Validation of the algorithm's robustness in identifying interleaving and hybrid patterns.

Main Results:

• The developed algorithm demonstrated high accuracy in detecting wandering patterns.
• Achieved an identification accuracy of 90% on a real-world dataset.
• The grid-based strategy proved effective for both indoor and outdoor scenarios.

Conclusions:

• The proposed unified algorithm offers a robust solution for indoor and outdoor wandering pattern detection.
• This approach facilitates a more comprehensive understanding of wandering behavior and its clinical implications.
• The novel grid-based representation is a key innovation enabling dual-environment applicability.