Simulation and Real-World Model Evaluation

Overview

This project performs simulations and evaluates model performance on both synthetic and real-world data. Follow the steps below to run the analysis.

Steps

1. Install Dependencies
   Run install_dependencies.R to install all the required libraries.

2. Run Simulation
   Execute simulation.R to:

   • Estimate model coefficients (betas)
   • Compute mean squared error (MSE) for both training and test datasets
   • Save the results for further analysis

3. Generate Plots
   Run plots.R to visualize:

   • Boxplots of the mean squared error (MSE)
   • Estimated covariates across models

4. Evaluate Model Performance
   Run table_TPR_NPR.R to calculate the following performance metrics:

   • True Positive Rate (TPR)
   • True Negative Rate (TNR)
   • Positive Predictive Value (PPV)
   • Negative Predictive Value (NPV)

   The results will be saved as table_TPR_NPR.RData.

5. Apply to Real-World Data
   Run real-life-study.R to apply the analysis on the real dataset Clean_Dataset.csv.