Machine Learning Models for Prediction and Risk Mapping of Zoonotic Disease Outbreaks in Southern States of India

Zoonotic diseases, which are transmitted between animals and humans, present a significant and growing public health concern, particularly in regions where human-animal-environment interactions are intense and dynamic. The Southern States of India-characterized by diverse climatic conditions, dense livestock populations, and complex socio-ecological systems-are particularly vulnerable to outbreaks of zoonotic diseases such as anthrax and enterotoxaemia. Addressing these challenges requires a multidisciplinary approach that combines traditional epidemiological understanding with advanced data science techniques, including machine learning and spatial statistics.

This book is an outcome of academic and applied research aimed at bridging the gap between veterinary science, geospatial technologies, and computational intelligence. The primary objective is to provide researchers, policymakers, and practitioners with a comprehensive guide to predictive modeling and spatial risk assessment of zoonotic disease outbreaks using data-driven approaches. The foundational understanding of zoonotic diseases, focusing on their classification, global impact, and relevance in the Indian context. Special emphasis is placed on anthrax and enterotoxaemia, exploring their epidemiology, clinical features, and preventive strategies. Following this, the text introduces the data-centric components necessary for predictive modeling, including outbreak records, remote sensing datasets, and meteorological variables. It delves into the importance of accurate data annotation and outlines the use of environmental parameters for forecasting disease risks.

Our journey begins with an exploration of machine learning models in disease prediction and with classification algorithms such as Random Forest, Support Vector Machine, and Adaptive Boosting. Model performance evaluation using metrics like the Kappa statistic and ROC curves is also discussed. Further, the application of spatial statistical methods for identifying disease hotspots and temporal trends is examined. Tools like the Getis-Ord Index, SaTScan, and the estimation of the Basic Reproduction Number (R₀) are covered in depth. The text culminates with the development of predictive risk maps and model-based risk assessments, providing practical insights into disease surveillance and control strategies.

We envision this book as a valuable resource for graduate students, public health researchers, veterinary epidemiologists, GIS professionals, and decision-makers involved in disease control and prevention. By integrating spatial science and machine learning, this work advocates for a proactive, data-informed approach to managing zoonotic disease threats in India and similar epidemiological settings worldwide.

We hope this book not only contributes to academic knowledge but also inspires further innovation in predictive epidemiology and spatial disease intelligence.