Consistency Evaluation of Land Use Distribution Pattern Supported by Spatial Autocorrelation

  Objectives  The consistency of spatial distribution pattern before and after generalization is an important factor to evaluate the quality of land use data. Considering the lack of quantitative analysis and visualization in spatial distribution model evaluation of data, we put forward a new evaluation method based on the unique autocorrelation of spatial data.

  Methods  Firstly, we establish the spatial weight matrix using the semantic distance of data and calculate global and local autocorrelation of data before and after processing by Moran's I index. Secondly, we use Moran scatter plot and LISA (local indicators of spatial association) aggregation map to visualize the quality of spatial distribution patterns.

  Results  Compared with the traditional data quality evaluation method, the proposed method can better evaluate the consistency of spatial distribution pattern of global map by taking into account the semantic relation of the data, obtaining the quantifiable aggregation degree and contrasting by the visual method.

  Conclusions  In cognitive experiments, we take the land type data been scale transformed which cause data quality problems mostly as an example. The experimental results accord with human cognitions and the experimental method is practical.