How can machine learning predict cholera: insights from experiments and design science for action research.

Area of Science:

• Public Health
• Infectious Disease Epidemiology
• Data Science

Background:

• Cholera is a significant cause of mortality in Nigeria, exacerbated by poor access to clean water and sanitation.
• Factors like natural disasters, illiteracy, and internal conflicts contribute to cholera's spread, particularly in refugee settings.
• Predictive modeling is crucial for mitigating cholera's impact in vulnerable regions.

Purpose of the Study:

• To develop and validate a cholera outbreak risk prediction (CORP) model utilizing machine learning.
• To enhance cholera outbreak detection and forecasting capabilities in Nigeria.
• To provide a data-driven tool for public health interventions.

Main Methods:

• Employed design science principles and machine learning for cholera outbreak prediction.
• Utilized Nonnegative Matrix Factorization (NMF) for dimensionality reduction and Synthetic Minority Oversampling Technique (SMOTE) for data balancing.
• Applied Density-Based Spatial Clustering of Applications with Noise (DBSCAN) for outlier removal and Extreme Gradient Boost for predictive modeling.

Main Results:

• The developed CORP model achieved a high accuracy of 99.62%.
• Demonstrated a Matthews's correlation coefficient of 0.976 and an Area Under the Curve (AUC) of 99.2%.
• Performance metrics showed significant improvement compared to previous studies.

Conclusions:

• The advanced CORP model offers a highly accurate solution for predicting cholera outbreaks in Nigeria.
• This machine learning-based approach can significantly aid healthcare providers in proactive cholera management.
• The model's effectiveness underscores the potential of data science in public health surveillance and response.