RULIE : rule unification for learning information extraction

Abstract

In this paper we are presenting RULIE (Rule Unification for Learning Information Extraction), an adaptive information extraction algorithm which works by employing a hybrid technique of Rule Learning and Rule Unification in order to extract relevant information from all types of documents which can be found and used in the semantic web. This algorithm combines the techniques of the LP2 and the BWI algorithms for improved performance. In this paper we are also presenting the experimen- tal results of this algorithm and respective details of evaluation. This evaluation compares RULIE to other information extraction algorithms based on their respective performance measurements and in almost all cases RULIE outruns the other algorithms which are namely: LP2 , BWI, RAPIER, SRV and WHISK. This technique would aid current techniques of linked data which would eventually lead to fullier realisation of the semantic web.