Evaluation of Triple Indices in Retrieving Web Documents

Indexing is one of the important technique used to make searching and retrieval process more accurate and faster. In this paper, the indexing features of triple stores are investigated to see the pattern of semantic indexing that will improve the current information retrieval indexing. To do the evaluation, we choose the best triple stores, Allegro Graph. Allegro Graph require more than one triple indices pattern to be matched with the query. Therefore we do the evaluation of triple indices to determine which triple indices pattern gives for fast retrieval. During experiment, we use RDF data. Then we matched the triple indices pattern with the SPARQL query to retrieve the relevant data. As the result, it would take longer if the system contains unneeded triple indices which will waste CPU times and disk space. Removing unneeded triple indices pattern will provide for fast retrieval.