Spatial bootstrapped microeconometrics: forecasting for out-of-sample geo-locations in big data

Email: Katarzyna Kopczewska

Abstract

Spatial econometrics for the big data point geo-locations has a limited possibility of forecasting with a calibrated model for the new out-of-sample geo-points. This is because of spatial weights matrix W defined for in-sample observations only as well as the computational complexity for a huge W. This paper proposes the novel methodology which calibrates both space and model using bootstrap and tessellation. Bootstrapping enables the calibration of the econometric model without the need for estimation on the whole dataset. The best bootstrapped model is selected with Partitioning Around Medoids (PAM) algorithm, which classifies the regression coefficients jointly, in a non-independent manner. Tessellation for the points used in the selected best model to tiles allows for a representative division of space. New out-of-sample points are assigned to tiles and linked to the spatial weights matrix as a replacement for an original point. This efficient procedure supports the big data geo-located point data and makes feasible a usage of calibrated spatial models as a forecasting tool for out-of-sample data.

Organizzatrice: Marzia Freo