Logo Lanfrica

RejoiceMsiska/modeling-malaria-mortality-

Domain:

healthcare

Record type:

model
Creator:
Rej
Host:
The aim is to model and forecast Malaria Mortality in Mozambique. # modeling-malaria-mortality- The aim is to model and forecast Malaria Mortality in Mozambique. # Contents of README.md # Malaria Forecasting Model (R Version) This repository contains an R-based forecasting model for malaria deaths using INLA for Bayesian Generalized Additive Models. It is designed to be compatible with the Climate and Health Assessment Platform (Chap). ## Structure - `input/`: Contains sample data files (trainData.csv, futureClimateData.csv). - `train.r`: Script to train the model (dummy for INLA approach). - `predict.r`: Script to generate predictions using the trained model. - `isolated_run.r`: Script to run training and prediction locally. - `MLproject`: MLflow project file for Chap integration. ## Usage 1. Install renv: `install.packages('renv')` and run `renv::init()` to manage dependencies (INLA, dplyr, readr, lubridate). 2. Run locally: `Rscript isolated_run.r` 3. For Chap: Follow Chap documentation to evaluate the model. Uses docker image with INLA pre-installed. Note: The model assumes input data has columns: time_period, rainfall, mean_temperature, disease_cases (for train), location, population. Lagged features are computed internally. Fitting occurs in predict.r by appending future data.