Improvements of SHIRI Extraction and Alignment Processes on Semi Structured Documents

To move the web from a Web of documents to a Web of data commonly named semantic web, it is obviously necessary to extract objects as terms and named entities that support a substantial proportion of texts semantic. This step could be the basis for existing web annotation which can improve the precision in handling documents for applications. The rapid growing of the web make impossible doing this manually. Many extraction and annotation techniques are used to resolve the first and easyiest problem of information search which is finding documents containing the searched data. In this work we proposed an experimentation of improvements on a model that allow locating and extracting the more exact parts of documents that correspond to the user request. This work extends SHIRI
¹
which is an ontology-based system for integration of semi-structured documents related to a specific domain. The ontology is described by a set of concepts, relations and their properties. It also contains a lexical part. It relies on an automatic, unsupervised and ontology-driven approach for extraction, alignment and querying for semantic annotation of tagged semi-structured documents elements. In this paper we focus on experimenting of two major improvements we proposed in our previous works. Experiments on a real datasets show that these improvements increase greatly the recall and the returned answers are effectively more precise and ranked according to their precision.