Using linked data to interpret tables
Description: Vast amount of information is available in structured forms like spreadsheets, database relations, and tables found in documents and on the Web. In today’s talk I will describe an approach that uses linked data to interpret such tables and associate their components with nodes in a reference linked data collection. Our proposed framework assigns a class (i.e. type) to table columns, links table cells to entities, and inferred relations between columns to properties. The resulting interpretation can be used for a variety of tasks such as annotating tables for the Semantic Web, confirm existing facts in the linked data collection, propose new facts to be added in the linked data cloud, answer queries over tables and data integration. This talk will focus on using the table interpretation for annotating web tables. We implemented a prototype which uses DBpedia as the linked data collection and Wikitology for background knowledge. We evaluated its performance using a collection of tables from Google Squared, Wikipedia and the Web. I will also address on issues that we will be working in the semester.
Authors: Varish Mulwad
Date: September 14, 2010
Format: Microsoft PowerPoint (Need a reader? Get one here)
Number of downloads: 444
Access Control: Publicly Available
Available for download as