Attribute and Positional Error in GIS

In many spatial data applications, the events at a location are highly influenced by the events at neighboring locations. In fact, this natural inclination of a variable to exhibit similar values as a function of distance between the spatial locations at which it is being measured is known as spatial dependence. Spatial autocorrelation is used to measure this spatial dependence. If the variable exhibits a systematic pattern in its spatial distribution, it is said to be spatially autocorrelated. The existence and strength of such interdependence among values of a specific variable with reference to a spatial location can be quantified as a positive, zero, or negative spatial autocorrelation. Positive spatial autocorrelation indicates that similar values or properties tend to be collocated, while negative spatial autocorrelation indicates that dissimilar values or properties tend to be near each other. Random patterns indicate zero spatial autocorrelation since independent, identically distributed random data are invariant with regard to their spatial location.