Machine learning project with MICS microdata in R
The purpose of the repository is to demonstrate how to use caret to build classification models. This document contains the scripts for machine learning models that aim to predict the use of contraception in Thailand, Laos, and Mongolia with data from the Multiple Indicator Cluster Surveys (MICS) of the UNICEF. The datasets are individual survey responses, and user registrations are required to access the microdata. Therefore, the data are not included in this repository.
The goal is to predict use, a binary variable that indicates whether a women of reproductive age is currently using contraception. The project report is available here.
- Analysis with
caret:- 🔍 Just the code?
analysis.R - 🔍 R markdown?
analysis.Rmd - 🔍 Markdown?
analysis.md
- 🔍 Just the code?
- Analysis with
tidymodels: This is a very brief script to experiment thetidymodelsecosystem.
-
Project report of Predicting Use of Contraception in Asia with Machine Learning Algorithms, in which I discuss the overall predictive modelling workflow, including analysis strategy, model comparison, and limitations.
-
A simple tutorial on Create Predictive Classification Models in R with Caret with more detailed explanation on the
caretpackage.