Spherekit: Error Analysis

Cross-validation

Cross-validation refers to the process of removing one of the n observation points and using the remaining n-1 points to predict its value. This process is repeated at each data point; for each estimate, n-1 points are used. The interpolation error at each data point is the difference between its observed and predicted values. An option is available to interpolate the error field to a uniform grid (using a user-specified method). This has the effect of reducing spatial bias in the estimate of the error.
The output of a cross-validation is an error field that can be manipulated as any other field. The display of an error field includes the error field, a table of error summary statistics, a histogram, and a scatterplot. The overall error is computed either as the mean average error (MAE) or the root mean square error (RMSE).