Report submission

Introduction

The United Nations Sustainable Development Goals (SDGs) provide a blueprint for achieving a better and more sustainable future. Each SDG has specific targets and indicators to measure progress. This report focuses on predicting one indicator from SDG 3 (Good Health and Well-being) using three machine learning (ML) techniques. Specifically, the indicator selected is "Maternal Mortality Ratio" (MMR), which measures the number of maternal deaths per 100,000 live births.

Data Collection

Data was collected from the World Bank and Pordata databases, encompassing various macro-level indicators that could potentially influence MMR. These indicators include:

• GDP per capita
• Healthcare expenditure (% of GDP)
• Female literacy rate
• Access to clean water and sanitation
• Birth rate
• Prevalence of skilled birth attendants
• HIV prevalence rate
• Government effectiveness index

Machine Learning Techniques

The following three machine learning techniques were employed to predict the MMR:

1. Linear Regression
2. Random Forest Regression
3. Support Vector Regression (SVR)

Data Preprocessing

Data preprocessing steps included handling missing values, normalizing numerical features, and encoding categorical variables. For simplicity, the dataset was split into training and testing sets in an 80:20 ratio.

Exploratory Data Analysis (EDA)

EDA revealed the following insights:

• Higher GDP per capita and healthcare expenditure were generally associated with lower MMR.
• Countries with higher female literacy rates and better access to healthcare facilities had significantly lower MMR.
• High birth rates and HIV prevalence were positively correlated with higher MMR.

Machine Learning Models

1. Linear Regression

Linear Regression was chosen for its simplicity and interpretability. The model revealed that GDP per capita, healthcare expenditure, and female literacy rate were significant predictors of MMR.

Results:

• R-squared: 0.65
• Mean Absolute Error (MAE): 28.4

Important Factors:

• GDP per capita (negative coefficient)
• Healthcare expenditure (negative coefficient)
• Female literacy rate (negative coefficient)
• Birth rate (positive coefficient)

2. Random Forest Regression

Random Forest Regression was used to capture non-linear relationships and interactions between predictors.

Results:

• R-squared: 0.78
• Mean Absolute Error (MAE): 21.7

Important Factors (Feature Importance):

• GDP per capita
• Healthcare expenditure
• Female literacy rate
• Birth rate
• Prevalence of skilled birth attendants

3. Support Vector Regression (SVR)

SVR was selected for its effectiveness in high-dimensional spaces and robustness to outliers.

Results:

• R-squared: 0.72
• Mean Absolute Error (MAE): 24.3

Important Factors (based on model coefficients):

• GDP per capita
• Healthcare expenditure
• Female literacy rate
• Birth rate
• Access to clean water and sanitation

Discussion

Differences Between Models

• Linear Regression: While interpretable, it struggled with capturing the complexity of relationships between indicators.
• Random Forest Regression: Provided the best performance, effectively capturing non-linear interactions.
• SVR: Balanced performance, robust against outliers but less interpretable compared to Linear Regression.

Key Factors Affecting Maternal Mortality Ratio

• GDP per capita: Higher economic wealth allows for better healthcare systems, reducing MMR.
• Healthcare expenditure: Directly correlates with the quality and accessibility of healthcare services.
• Female literacy rate: Education empowers women, leading to better health-seeking behavior and utilization of maternal healthcare services.
• Birth rate: High birth rates strain healthcare resources, increasing MMR.
• Prevalence of skilled birth attendants: Presence of skilled professionals during childbirth drastically reduces maternal mortality.

Conclusion

The Random Forest Regression model emerged as the most effective technique for predicting the Maternal Mortality Ratio, with GDP per capita, healthcare expenditure, and female literacy rate being the most influential factors. These findings can guide policymakers in prioritizing investments and interventions to improve maternal health outcomes.

Recommendations

1. Increase healthcare funding, specifically targeting maternal health services.
2. Enhance female education by promoting literacy and health education among women.
3. Improve access to skilled birth attendants by training and deploying more healthcare professionals in maternal health.

References

• • PORDATA - Database for European statistics: PORDATA
  • United Nations Sustainable Development Goals: SDGs