Predictive risk mapping of human leptospirosis using support vector machine classification and multilayer perceptron...

Area of Science:

• Epidemiology
• Spatial Analysis
• Machine Learning

Background:

• Leptospirosis is a zoonotic disease linked to contaminated water and environments.
• Increasing cases in northern Iran necessitate risk identification for prevention.
• Understanding disease distribution is crucial for public health interventions.

Purpose of the Study:

• To develop predictive risk maps for leptospirosis in northern Iran.
• To identify spatial patterns and environmental factors associated with leptospirosis incidence.
• To evaluate machine learning models for leptospirosis risk prediction.

Main Methods:

• Spatial statistics (Moran's I) to analyze disease clustering.
• Pearson correlation to assess relationships between environmental factors and leptospirosis.
• Machine learning algorithms (Support Vector Machine, Multilayer Perceptron) for risk mapping.

Main Results:

• Leptospirosis cases exhibited a highly clustered distribution in the study area.
• Environmental and topographical factors showed significant correlations with disease incidence.
• Both machine learning models demonstrated adequate performance in predicting leptospirosis risk.

Conclusions:

• Predictive risk maps are valuable tools for public health policy and leptospirosis control.
• Machine learning approaches effectively model complex spatial-temporal patterns of zoonotic diseases.
• Spatial analysis combined with machine learning offers a robust strategy for disease surveillance.