An Automatic Semantic Extraction Method for Web Data Interchange

Data interchange in Internet is mainly for specific web applications. The data formats are simple but lack of rich semantic descriptions, which can't meet current requirements of deep web data analyzing with semantic technology. Aiming at connecting web data with semantics, we propose an automatic semantic extraction method to handle web data sets with semantics and generate semantic data for applications, which includes data parsing, semantic mapping, semantic enrichment, and ontology merging processes. This method converts web data into semantic web descriptions and improves data semantics according to semantic computation. Meanwhile, it builds semantic models for data instances, which can be applied to further semantic reasoning applications. We use this method to extract schemaless JSON data automatically, including concepts, properties, constrains and values, and build semantic ontology to describe the metadata and instances. The experimental results show that this method can process web data resources and create semantic data effectively.