A statistical simulation study investigating estimation techniques for the Three-Parameter Generalized Exponential Distribution under non-regular likelihood conditions.
This project evaluates several estimation methods for the Three-Parameter Generalized Exponential Distribution (TPGED), where estimation becomes challenging due to an unknown location parameter and non-regular likelihood behavior.
Implemented approaches include:
- Maximum Product of Spacings (MPS)
- Modified Maximum Likelihood Estimation (MMLE)
- Location Scale Free Likelihood (LSPF)
The project combines theoretical derivations, numerical optimization, Monte Carlo simulation, and real-world reliability data analysis.
- Statistical estimation under non-regular models
- Monte Carlo simulation studies
- Bootstrap confidence intervals
- Profile likelihood analysis
- Reliability modelling
- Goodness-of-fit evaluation
- Numerical optimization using L-BFGS-B
- Publication-quality plots generated using Python
three-parameter-ged-simulation/
├── ged_estimation_simulation.ipynb ├── report/ │ └── three_parameter_ged_report.pdf ├── presentation/ │ └── presentation.pdf ├── latex/ │ ├── main.tex │ ├── references.bib │ ├── report_assets/ │ ├── contour_plots.png │ └── plot.png └── README.md
- Python
- NumPy
- SciPy
- Pandas
- Matplotlib
- Jupyter Notebook
- LaTeX
- Trisha Mukherjee
- Aayush Aneja
Supervisor: Dr. Suchismita Das