ENSM-SE at INEX 2012: Basic Experiments

Our objective in the INEX 2012 campaign was to integrate the semantic tags and the linked data in our proximity retrieval model. This model was sucessfully used in previous INEX campaigns and obtained good results, particularly in 2007 with the second place in the Ad Hoc Track Focused Task [1], and in 2010 with the rst place in the Ad Hoc Track Relevant in Context Task [2] Though we had several discom tures with the collection because i) there were several versions of the collection, the last one available at the end of June, one week before the initial run submission deadline, ii) the di erent versions were di cult to follow because they were not clearly identi ed, iii) not every documents were well formed according to the XML format, iv) the provided DTD gives little information on the actual structure and its semantics, v) the documents contains many semantic annotations but the underlying ideas used to generate them are not documented making them di cult to apprehend. We present in section 2 how we processed the documents to alleviate the problems with the DTD. Thus we only have been able to do some basic experiments presented in section 3. In section 4 we present our work in progress.