Automated Linear Geometric Conflation for Spatial Data Warehouse Integration Process
L. Savary, Karine Zeitouni · 2005
In spatial data warehouses, the quality of data greatly depends on the integration and cleaning process during which the data warehouse is fed. In this paper, we present a new geometric conflation algorithm for linear data type which gives better quality results than existing algorithm and thus, improves the quality of spatial data warehouses. While most of existing methods focus either on metric distance or shape to match spatial data, our algorithm takes into account these two factors. Consequently it resolves many complex cases which are not resolved by existing methods, and limits the need of expert intervention. In fact, the returned result, using the proposed algorithm, is close to human visual intuition. Experimental results show the effectiveness of our algorithm.