A Comparison Study on Criteria to Select the Most Adequate Weighting Matrix

Area of Science:

• Econometrics
• Spatial Analysis
• Statistical Modeling

Background:

• Spatial econometrics relies heavily on user-defined spatial weighting matrices (W).
• The selection of W is often subjective, based on prior knowledge, presenting the
• W issue
• .

Purpose of the Study:

• To develop and evaluate a new, objective method for selecting the optimal spatial weighting matrix (W) from a set of potential matrices.
• To address the limitations of the traditional aprioristic approach in spatial econometrics.

Main Methods:

• A novel method based on the entropy of the probability distribution estimated from the data is proposed for W matrix selection.
• The study reviews existing common alternatives for W matrix selection.
• A large Monte Carlo simulation is conducted to assess the proposed method's effectiveness.

Main Results:

• The proposed entropy-based method demonstrates effectiveness in selecting appropriate spatial weighting matrices.
• Comparative analysis through Monte Carlo simulations validates the performance of the new approach against existing methods.
• A case study illustrates the practical application of the method.

Conclusions:

• The entropy-based approach offers a data-driven and objective solution to the spatial weighting matrix selection problem.
• This method provides a valuable alternative to subjective user-defined matrices in spatial econometrics.
• The findings suggest improved reliability and objectivity in spatial econometric modeling.