Predictive Modeling of Typhoid Incidence in Nepal

Master’s Thesis — four-model framework with Weighted Ensemble

Thesis · Master’s · About

About	Abstract · Highlights · Figures · Read the paper · How to cite
Paper (APA)	Landing · Introduction · Literature Review · Methodology · Results · Discussion · Data · Sources
Presentation	Slides with APA in-text citations & References
Notebook	Python notebook — runs thesis/main.py with supporting modules

Abstract

Typhoid fever, caused by Salmonella enterica serovar Typhi, imposes a persistent and disproportionate public-health burden on Nepal, particularly among communities with inadequate access to safe water, improved sanitation, and primary healthcare. This thesis develops, trains, and rigorously evaluates four machine-learning predictive models — Random Forest (RF), Extreme Gradient Boosting (XGBoost), Multilayer Perceptron (MLP), and a Weighted Ensemble — for the prediction of monthly district-level typhoid case counts across Nepal for the period January 2015 to December 2023 (108 months; N = 7,327 district-month observations). Four validated primary data sources are integrated: HMIS outpatient enteric-fever surveillance, DRR-portal flood records, CHIRPS precipitation, and ERA5-Land temperature and humidity.

The Weighted Ensemble achieves the strongest predictive performance across all evaluation criteria: R² = 0.8675, RMSE = 126.20 cases per district per month, MAE = 68.07 cases, MAPE = 41.01 %. Individual model performance is competitive (XGBoost R² = 0.8614, MLP R² = 0.8635, RF R² = 0.8563). Feature-importance analysis identifies the prior month’s typhoid case count (cases_lag1) as the single most influential predictor (normalized importance = 0.670), followed by one-month lagged mean temperature (0.088), cyclical month encoding (combined: 0.122), and the monsoon indicator (0.037). The study offers actionable evidence for the development of climate-informed typhoid early-warning systems aligned with Nepal’s National Adaptation Plan 2021–2050.

Highlights

0.87 Weighted Ensemble R² · held-out 12 months

126 Test-set RMSE · cases / district-month

7,327 District-month observations

3 × August monsoon peak vs. dry-season median

• First national-scale, district-level, multi-model ML study of typhoid fever in Nepal — all 77 districts, nine years of HMIS surveillance.
• R² = 0.8675 on a strictly chronological 2022–2023 held-out test set — at or above international benchmarks for comparable climate-health ML studies.
• Weighted Ensemble outperforms all three individual component models across all four evaluation metrics (RMSE, MAE, R², MAPE).
• August peak is ~3× the dry-season baseline; Madhesh Province carries the heaviest burden, driven by low WASH coverage and high flood exposure.
• End-to-end reproducible — every number is regenerable by running thesis/main.py.

Figures

The five publication figures generated by thesis/main.py are committed to thesis/output/figures/:

#	Figure
4.1	National monthly typhoid time series (2015–2023)
4.2	Pearson correlation matrix — climate × cases
4.3	XGBoost gain-based feature importance
4.4	Actual vs. predicted scatter plots — four-model comparison
4.5	Seasonal distribution of district-monthly case counts

Read the paper

• Formal thesis (PDF) — thesis.docx.pdf · the complete five-chapter scholarly submission to Kathmandu University.
• APA paper version — seven-section Paper page: Introduction · Literature Review · Methodology · Results · Discussion · Data · Sources.
• Slide deck — Presentation with APA in-text citations and a References slide.
• Code walkthrough — Notebook with main.py and all seven supporting modules embedded inline.

How to cite

Baral, K. (2026). Predictive modeling of typhoid incidence in Nepal under extreme climate change scenarios using machine learning [Master’s thesis, Kathmandu University]. https://baralsamrat.github.io/publications/thesis/

@mastersthesis{baral2026thesis,
  author  = {Baral, Kritika},
  title   = {Predictive Modeling of Typhoid Incidence in Nepal
             Under Extreme Climate Change Scenarios Using
             Machine Learning},
  school  = {Department of Health Informatics, Kathmandu University},
  address = {Dhulikhel, Nepal},
  year    = {2026},
  url     = {https://baralsamrat.github.io/publications/thesis/}
}