Semplore

Semplore is a semantic search engine towards providing hybrid query capability and scalability for querying and searching semantic web data using existing IR engines.

1. Semplore: An IR Approach to Scalable Hybrid Query of Semantic Web Data, ISWC+ASWC 2007 Spotlight Paper

2. ISWC+ASWC 2007 Presentation Chart

3. Wikipedia dataset

Wikipedia is used as the real world knowledge base, which contains rich textual descriptions and relationships between entities. We combine TBox from YAGO and ABox from DBpedia as our Wikipedia dataset. Some simple heuristic rules are used for data cleaning, such as case unification, predicate stemming, and hyphen unification.

download here

4. Query set

Since there are no benchmarks for hybrid queries yet, we created four sets of queries for the Wikipedia dataset with increasing complexity in query patterns, from node, path to tree. The queries include 10 TBox queries and 40 ABox queries. All the queries contain one or more keyword searches to express vague information needs.

download here

5. RelSE: Relational Search on wikipedia powered by Semplore

6. A simple faceted search powered by Semplore

The flash shows an example of a simple faceted search powered by Semplore on the Wikipedia dataset. Have a try on the online demo system.