The effect of spatial autocorrelation on spatial co-location pattern mining

spatial co-location pattern mining is an important part of spatial data mining, and the purpose is to discover the coexistence spatial feature sets whose instances are frequently located together in a geographic space. However, it ignores the existence of autocorrelation features that is not associated with surrounding features. For example, “cactus” and “jerusalem artichoke” are two common plants in the desert, and it is easy to get prevalent pattern {cactus, jerusalem artichoke} for the existing spatial co-location pattern mining frameworks, but biologists have determined that “jerusalem artichoke” is a spatial autocorrelation feature so that above pattern is meaningless. To avoid getting prevalent patterns that contain spatial autocorrelation features, we propose an algorithm to find and remove the spatial autocorrelation feature from spatial data sets, so that we can get really meaningful prevalent co-location patterns, and the experimental results over synthetic/real data sets show the effectiveness and feasibility of our method.