Research And Application Of A Query Index For SDM In GIS

What spatial data mining studies is the spatial relationship between subject objects and objects in their neighbors. Spatial data mining needs to query various spatial relations include topology relation, direction relation, distance relation and any their combination. These characteristics of spatial data mining query brought challenge to spatial index and data access technology at present. To query spatial relation, we must carry out spatial connection operation. If being to query various spatial relations, we need to carry out spatial connection operation many times. It will cause great amount process cost that uses existing spatial index and data access to support spatial data mining query, since spatial connection is the most complex and expensive operation in spatial database. Owing to the above thought, this thesis aims at the problem for spatial data mining query efficiency and the following work are done:An efficient method for processing direction relation queries using R-trees is presented. The quad-tuples model is defined to represent direction relations between MBRs of spatial objects. An algorithm of processing the filter step using R-trees is given and the refinement step is further decomposed into three different operations. Experiment results show that the proposed method can efficiently process direction relation queries between objects of any data types in 2D space.A method based on the incremental implementation of Distance-associated Join Index (DJI) is presented to efficiently process range distance join queries in OLAM environments. The proposed method uses selectivity estimation approaches of distance join to selectively materialize part of DJI. To efficiently compute the DJI, it also uses different distance functions to fast prune when traversing R-trees. Experiment results show that the proposed method realizes a trade of between computation and storage costs, and it is an efficient method for processing on-line range distance join queries.The indexing and data access support to realize efficient SDM query execution is studied. The Distance-Direction associated Join Indices (DDJI) and its incremental implementation method is presented, and the unified implementation approach of SDM algorithms based on DDJI is studied. Based on DDJI, a SDM experimental system is designed and implemented. The above achievements are validated in the system and the experiment results show that they are effective and practical.