|
Comparing the performance of machine learning algorithms in predicting the spatial distribution of cutaneous leishmaniasis using a dual-map approach
|
Zeinab Neisani Samani   , Ali Asghar Alesheikh * |
|
|
|
Abstract: (15 Views) |
Cutaneous leishmaniasis, a common disease of humans and animals, remains one of the persistent public health challenges in endemic areas. In this study, by integrating spatial information systems and machine learning algorithms, we investigated the impact of environmental and spatial factors on the distribution pattern of cutaneous leishmaniasis in Ilam Province (western Iran) during 2014–2019. Disease incidence data were combined with climatic variables. To overcome the limitation of presence-only data, a comparative modeling framework was developed by generating pseudo-absence data and producing spatial dual maps. Three algorithms, support vector machine, random forest, and logistic regression, were implemented. The random forest model demonstrated superior performance compared to the other models, achieving evaluation metrics including an AUC-ROC of 0.9995, a recall of 0.92, a precision of 0.88, an F1-score of 0.90, and an accuracy of 0.9988. Feature importance analysis identified maximum mean temperature (TMax_M) as the most influential predictor variable. The output maps showed that high-risk hotspots were mainly concentrated in the central and southwestern regions of the province. The spatial findings of this study highlight the critical relationship between specific climatic drivers and disease hotspots. By providing dual maps (probability and risk), this study offers valuable practical evidence for health authorities to prioritize surveillance and preventive measures. |
Article number: 1 |
| Keywords: Cutaneous leishmaniasis, machine learning, geospatial health, spatial modeling, dual-map framework, climatic factors. |
|
|
|
Type of Study: Research |
Subject:
GIS
Received: 2025/11/3 | Accepted: 2025/12/8
|
|
|
|
|
|
|
| Send email to the article author |
|
|
|
| Add your comments about this article |
|
|
|
