Missing Value Imputation for Multivariate Data using Data Depth

M.Sc. Statistics Academic Project
Central University of Rajasthan

---

📌 Project Overview

This academic project addresses the challenge of missing values in multivariate datasets by leveraging data depth–based statistical techniques. The aim is to develop robust imputation methods that preserve the intrinsic structure and statistical properties of data.

🧠 Methodology

• Developed data depth–based imputation techniques for multivariate data
• Simulated missingness levels ranging from 5% to 15%
• Benchmarked against standard imputation methods:
  • KNN Imputation
  • Random Forest Imputation
  • Mean Imputation
• Evaluated performance using error-based metrics and bias–variance analysis

🛠 Tools & Technologies

• Programming Language: R
• Statistical Techniques: Multivariate Analysis, Data Depth Methods

📂 Repository Contents

• Final Thesis (PDF)
• Project Report (PDF)
• Final Presentation (PDF)

🚀 Future Enhancements

The repository will be updated with R scripts, simulation code, and extended analysis in future revisions.

---

Submitted in partial fulfillment of the requirements for the M.Sc. in Statistics.