A labeled-tree approach to semantic and structural data interoperability applied in hydrology domain

作者：Ravindran Nimmy; Liang Yao^*; Liang Xu

来源：Information Sciences, 2010, 180(24): 5008-5028.

DOI：10.1016/j.ins.2010.06.015

• Data integration
• Unstructured and semi-structured data
• Semantic and structural heterogeneities
• Autonomous data sources
• Metadata driven

The issues of data integration and interoperability pose significant challenges in scientific hydrological and environmental studies, due largely to the inherent semantic and structural heterogeneities of massive datasets and non-uniform autonomous data sources. To address these data integration challenges, we propose a unified data integration framework, called Hydrological Integrated Data Environment (HIDE). HIDE is based on a labeled-tree data integration model referred to as Datallode tree. Using this framework, characteristics of datasets gathered from diverse data sources - with different logical and access organizations - can be extracted and classified as Time-Space-Attribute (TSA) labels and are subsequently arranged in a Datallode tree. The uniqueness of our approach is that it effectively combines the semantic aspects of the scientific domain with diverse datasets having different logical organizations to form a unified view. Further, we also adopt a metadata-based approach for specifying the TSA-Datallode tree in order to achieve flexibility and extensibility. The search engine of our HIDE prototype system evaluates a simple user query systematically on the TSA-Datallode tree, presenting integrated results in a standardized format that facilitates both effective and efficient data integration.

• 出版日期2010-12-15
• 单位美国弗吉尼亚理工大学(Virginia Tech)