Nonparametric Density Estimation for Data Scattered on Irregular Spatial Domains: A Likelihood-Based Approach Using...

Area of Science:

• Spatial Statistics
• Nonparametric Statistics
• Computational Geometry

Background:

• Accurate data density estimation is vital for informed decision-making and modeling.
• Existing methods struggle with data on irregular spatial domains.

Purpose of the Study:

• To develop a novel nonparametric density estimation procedure for data on irregular spatial domains.
• To provide theoretical guarantees for the proposed method's convergence.

Main Methods:

• Utilizing bivariate penalized spline smoothing over triangulation.
• Employing a likelihood-based approach with a regularization term for the logarithm of density.
• Incorporating a second-order differential operator to address density roughness.

Main Results:

• Established asymptotic convergence rates in L2 and L-infinity norms under mild conditions.
• Demonstrated superior efficiency, flexibility, smoothness, and continuity compared to existing techniques.
• Validated through simulations and application to real-world motor vehicle theft data.

Conclusions:

• The proposed method offers a robust and effective solution for density estimation on irregular spatial domains.
• The technique provides enhanced accuracy and theoretical underpinnings for spatial data analysis.