Development of a Dog Health Score Using an Artificial Intelligence Disease Prediction Algorithm Based on Multifaceted...

Area of Science:

• Veterinary Medicine
• Artificial Intelligence
• Animal Behavior

Background:

• Canine diseases can manifest as aberrant behaviors or altered owner-dog interactions.
• Dog owners often lack reliable methods to independently assess their pet's health status.
• There is a need for accessible tools to monitor canine well-being using everyday data.

Purpose of the Study:

• To develop a methodology for dog owners to evaluate canine health using daily behavioral data.
• To create an AI model that quantifies dog behavior into a quantifiable 'Health Score'.

Main Methods:

• Collected three months of individual canine activity data (scratching, licking, swallowing, sleeping).
• Trained an AI model to identify abnormal behaviors and assign a numerical 'Health Score' (1-10).
• Validated the Health Score against veterinary diagnoses.

Main Results:

• The AI model generates a 'Health Score' from daily behavioral data.
• A Health Score below 5 suggests the need for veterinary consultation; scores above 5 indicate a healthy state.
• The Health Score achieved an 87.5% concordance rate with veterinarian diagnoses.

Conclusions:

• The developed Health Score reliably assesses canine health through continuous monitoring of daily activities.
• This AI-driven approach empowers dog owners to better understand and manage their pet's health.
• The Health Score offers a significant benefit for proactive pet healthcare management.