Imputing-Missing-Reaction-Times-Using-the-EM-Algorithm-Under-a-Bivariate-Normal-Model

This repository contains an R Markdown analysis where missing reaction time data for two stimuli (Stimulus A and Stimulus B) is imputed using the Expectation-Maximization (EM) algorithm under the assumption of a bivariate normal distribution.

📁 Project Overview

• 📄 Title: Imputing Missng Reaction Times Using the EM Algorithm
• 🧪 Method: Expectation-Maximization (EM) Algorithm
• 📊 Data: Reaction time dataset from Duke University

🛠️ Features

• Custom EM algorithm implementation in R
• Imputes values based on conditional expectations under multivariate normality
• Tracks convergence through log-likelihood evaluation
• Compares mean and variance before and after imputation
• Preserves correlation structure between variables
• Comprehensive visualizations including:
  • Histograms
  • Boxplots
  • Scatter plots (before and after imputation)

📦 Dependencies

Make sure the following R libraries are installed:

library(ggplot2)
library(gridExtra)
library(dplyr)

##local dev
1. open README.md on your local