面向数据集成的基于规则的自动语义转换

Data Integration Oriented Automatic Semantic Translation Based on Rules

为了定位对应于同一个实例在不同模式下的元组,提出了逻辑主键与全字段匹配相结合的定位算法,该方法是AuSTR解决语义冲突问题的技术核心,通过定义逻辑主健可以实现大部分数据行的定位,其他情形则采用全字段匹配算法,二者结合能对不同模式下的数据进行语义一致的定位,自动完成异构系统数据集成中的语义转换.AuSTR在数字校园中的应用证明了该方法能有效地对实时集中的增量数据自动完成语义转换.

In order to locate tuples which correspond to one instance in different schema, the authors propose a location algorithm imposing a logic primary key and the whole fields match. Meanwhile it is the core of AuSTR to resolve semantic conflicts. It can locate most of tuples by defining the logic primary key, while the other case adopts the whole fields match method. Combination of these two parts can locate data according to semantics under different schema,so that semantic translation for data integration of heterogeneous system can be accomplished automatically. The successful application of AuSTR to a digital campus project verifies that during real time collection the AuSTR is valid to fulfill semantic translation of real time increment data automatically.