Status : Verified
| Personal Name | Marcelo, Ronniel; Bayutas, Benjamin |
|---|---|
| Resource Title | Predicting and Understanding Zero-Dose Children in the Philippines: A Machine Learning Approach |
| Date Issued | June 2026 |
| Abstract | This project identified factors associated with zero-dose status among children in the Philippines using pooled data from five waves of the Philippine Demographic and Health Survey. Zero-dose children, defined as children who have not received any dose of a diphtheria-tetanus-pertussis-containing vaccine, are an important equity indicator under the Immunization Agenda 2030 because they represent children who may lack access to, or may have never been reached by, routine immunization services. Several machine learning models were developed and evaluated to predict whether a child aged 12–23 months was zero-dose. The final selected model was a neural network using LASSO-selected predictors and class-cost weighting, with one hidden unit, weight decay of 0.10, 32 input features, and a decision threshold of 0.41. This model was selected based on its ability to balance sensitivity and specificity in identifying children at risk of being zero-dose. SHAP-based model interpretability analysis identified survey year, religious affiliation, maternal education, place of delivery, household wealth, and postnatal care utilization as the most influential predictors of zero-dose status. The results showed that toddlers from poorer households, those born at home, those lacking postnatal care, and those whose mothers had lower educational attainment generally exhibited higher predicted probabilities of being zero-dose. These findings provide evidence on the important roles of socioeconomic conditions, healthcare access, and temporal factors in zero-dose status among Filipino children and may support more targeted planning for immunization programs aligned with the equity goals of Immunization Agenda 2030. |
| Degree Course | Professional Master in Data Science (Analytics) |
| Language | English |
| Keyword | Zero-dose children, immunization coverage, machine learning, predictive modeling, SHAP analysis, Demographic and Health Survey (DHS), vaccine equity, public health analytics, Immunization Agenda 2030 (IA2030) |
| Material Type | Thesis/Dissertation |
Preliminary Pages
7.88 Mb
Category : F - Regular work, i.e., it has no patentable invention or creation, the author does not wish for personal publication, there is no confidential information.
Access Permission : Open Access