Published April 29, 2025 | Version 1.0

Modeling Implied Volatility Surfaces using 2D Penalized B-Splines: A Numerical Analysis

  • 1. Computational Mathematics Program, University of Colorado Boulder

Description

This report investigates numerical methods for constructing implied volatility surfaces (IVS) from financial option data, with a focus on stability, convergence, and error analysis. Developed as a capstone for the Computational Mathematics Certificate at the University of Colorado Boulder, the study applies Newton-Raphson root-finding, natural cubic spline interpolation, and two-dimensional penalized B-splines solved with LSQR. The work explores the trade-off between accuracy and smoothness in volatility surface modeling and demonstrates that minimally penalized fits (λ ≈ 0.0001–1.0) achieve strong performance while avoiding oscillations. The findings contribute to applied numerical analysis in finance and illustrate practical applications of computational mathematics techniques.

Files

Report.pdf

Files (8.9 MB)

Name Size
md5:0cf93a18fb5b7ecadce8111cfa61333e
8.9 MB Preview Download

Additional details

Related works

Is supplement to
Dataset: https://home.treasury.gov/ (URL)
Is supplemented by
Dataset: https://finance.yahoo.com/ (URL)

Dates

Accepted
2025-05-05

References

  • Burden, R. L., Faires, J. D. (2011). Numerical Analysis (9th ed.). Brooks/Cole, Cengage Learning.
  • Paige, C. C., Saunders, M. A. (1982). LSQR: An algorithm for sparse linear equations and sparse least squares. ACM Transactions on Mathematical Software (TOMS), 8(1), 43–71.
  • Craven, P., Wahba, G. (1979). Smoothing noisy data with spline functions: estimating the optimal value of the smoothing parameter. Numerische Mathematik, 31(4), 377–403.