Determine Weights Of Evidence Correlation


This container calculates the correlation between a pair of maps.


Name Type Description
Ranges Weights Pre-defined intervals for continuous gray-tone variable.
Transition Transition Set A specified transition.

Optional Inputs


Name Type Description
Report Table Table containing a full report of correlation calculation. These are essentially the same results showed in the message log, but in a table format.



When considering the advantages of Weights of Evidence over other statistical methods, such as Logist or Linear Regression, one can mention that it is not constrained by the classical assumptions of parametric methods, which spatial data often violate. Furthermore, the effect of each spatial variable can be calculated independently of a combined solution. The only assumption that must be made is that the predictive maps are spatially independent, which can be tested using pairwise tests for categorical maps, such as the Cramers Coefficient, Contingency Coefficient, and the Joint Information Uncertainty (Bonham-Carter, 1994). As a result of these tests, one of the correlated variables can be either eliminated or combined with the second to form a new variable that will replace both in the integrated model. The two former methods are based on the chi-square statistic while the latter is derived from the Joint Entropy measure; all methods are calculated from a contingency table produced by cross-tabulating pair of maps.


Bonham-Carter, G 1994. Geographic information systems for geoscientists: modelling with GIS. New York, Pergamon.

Internal Name