A MULTI-LEVEL SPATIAL ASSOCIATION RULES MINING ALGORITHM BASED ON QUALITATIVE SPATIAL REASONING

摘要

Many Spatial relation models applied for spatial data mining do not have powerful representation ability, are hard to be understood or can not support uncertainty. A multi-level spatial relation model for approximate regions in spatial database Is proposed based on Fuzzy Logic and the RCC theory of Qualitative Spatial Reasoning. The mining multi-level spatial association rules algorithm QSRSAR is then put forward, QSRSAR is optimized by MBR prior judgment and node approximation. The experiment showsQSRSAR is better than other similar algorithms in expression, understandability and efficiency.