# AI-based analysis of climatic and air pollution determinants of dog bite incidence

## TL;DR

This study uses AI to analyze how climate and air pollution affect dog bite incidents in five Indian states, finding temperature and humidity as key factors.

## Contribution

The study introduces an AI model to predict dog bite risk based on environmental factors in diverse Indian regions.

## Key findings

- Higher maximum temperature is significantly associated with increased dog bite incidence.
- The AI model achieved 87% accuracy in predicting dog bite risk.
- Environmental factors alone do not fully explain regional variability in dog bite incidents.

## Abstract

Dog bite incidents are an emerging public health concern that may be influenced by changing environmental conditions. This study investigated the relationship between meteorological variables (maximum temperature and relative humidity) and dog bite incidence across five Indian states: Bihar, Karnataka, Punjab, Telangana, and Uttar Pradesh. The role of key air pollutants, including formaldehyde, nitrogen dioxide, sulfur dioxide, and ozone, was also examined. Statistical analyses showed that maximum temperature (p = 0.0014) and relative humidity (p = 0.0252) were significantly associated with dog bite incidence, with higher temperatures associated with increased incidence and higher humidity with reduced incidence. Principal component analysis (PCA) revealed no apparent clustering or dominant trend in environmental factors, indicating that temperature and humidity alone do not sufficiently explain dog bite variability across regions. Correlation analysis across monthly data demonstrated a strong overall positive association with maximum temperature (r = 0.84), although short-term annual trends show nonlinear fluctuations influenced by additional contextual factors. To predict dog bite risk, an artificial intelligence model (H2O XGBoost) was developed, achieving 87% accuracy and a mean absolute percentage error of 9.6%. This study highlights the importance of localized environmental interpretation and region-specific variability, contributing to understanding the ecological determinants of animal-related injuries and supports Sustainable Development Goals 3 (good health and well-being), 11 (sustainable cities and communities), and 13 (climate action) by informing strategies for safer and more resilient urban environments.