Entity Matching in Vector Spatial Data

Entity matching is a crucial and hard technology in application of vector spatial data integrating, data updating and map differential analyses. According to the disadvantage of matching algorithms nowadays, from the candidate searching algorithm of entity matching, similarity measure and matching strategy, this paper does a deep research in this three aspects. And proposes an area entity searching algorithm based on interior intersection, and another line entity searching algorithm based on buffer division. Both could enhance the seed of searching candidate match set, decreases the number of candidate matching entities. Also in this paper, giving an area entity integrated similarity measure index integrated entity area, overlap area, barycentre distance etc, and a line entity integrated similarity measure index based on buffer overlap area, azimuth code, length etc. . Both of them improve the recognizing ability of identical entity. Through the match strategy of bidirectional matching and clustering combination, effectively achieves the entity matching of one-to-many and many-to-many. ∗ Corresponding author: FU Zhongliang, Email: [email protected]