Two judging criteria to check validity of a model for filling gaps caused by incomplete geospatial data

Many models can be used to fill the gaps caused by incomplete geospatial data. But not all are valid. To study the validity of geospatial information diffusion model, in this article, two judging criteria are suggested to check if a model is valid for filling a gap unit. The root mean squared error of a model with a given sample after removing a test point is called datum error of the model. The error between real value and estimated value of the test point is called forecasting error of the model. The first criterion says that, when the average forecasting error is less than the average datum error, the model is invalid. The second criterion says that, the smaller the errors, the more valid the model. The results of computer simulation show that geospatial information diffusion model is more valid than the geographically weighted regression and the back propagation neural network.