Geospatial Information Diffusion Technology Supporting by Background Data

Abstract

In this paper, we express the initial concept of geospatial information diffusion supporting by background data, which plays a role as a bridge to diffuse the information carried by the observations, obtained from observed units, to gap units. The self-learning discrete regression, based on the multivariate normal diffusion, is suggested to supplement incomplete geospatial data to be complete. The suggested method has obvious advantages over the geographic weighted regression and the artificial neural network for inferring the observations in gap units