2 GIS空间特性与表达.ppt

Spatial Autocorrelation Non-spatial independence suggests many statistical tools and inferences are inappropriate. Correlation coefficients or ordinary least squares regressions (OLS) to predict a consequence assumes that the observations have been selected randomly. If the observations, however, are spatially clustered in some way, the estimates obtained from the correlation coefficient or OLS estimator will be biased and overly precise. They are biased because the areas with higher concentration of events will have a greater impact on the model estimate and they will overestimate precision because, since events tend to be concentrated, there are actually fewer number of independent observations than are being assumed. Indices of Spatial Autocorrelation Moran’s I Geary’s C Ripley’s K Join Count Analysis Moran’s I One of the oldest indicators of spatial autocorrelation (Moran, 1950). Still a defacto standard for determining spatial autocorrelation Applied to zones or points with continuous variables associated with them. Compares the value of the variable at any one location with the value at all other locations Moran’s I Where N is the number of casesXi is the variable value at a particular locationXj is the variable value at another locationX is the mean of the variableWij is a weight applied to the comparison between location i and location j Moran’s I Wij is a contiguity matrix If zone j is adjacent to zone i, the interaction receives a weight of 1 Another option is to make Wij a distance-based weight which is the inverse distance between locations I and j (1/dij) Compares the sum of the cross-products of values at different locations, two at a time weighted by the inverse of the distance between the locations Similar to correlation coefficient, it varies between –1.0 and + 1.0 When autocorrelation is high, the coefficient is high A high I value indicates positive autocorrelation Moran’s I Problems with weights Potential for distorted I value. Wij is normal