AOT / AOTL results for OAEI 2014

In this paper, we describe our two ontology alignment systems AOT and AOTL respectively. The AOT system uses different terminological matchers with a local filter and the AOTL system consists in combining the different terminological with linguistic matchers. The AOT and AOTL systems are designed for the ontology matching tracks in order to discover new semantic correspondences between entities of different ontologies to be aligned. This is the first participation of AOT and AOTL at OAEI 2014, we present the results obtained by running the first version of our systems in different tracks of OAEI 2014 evaluation campaign. 1 Presentation of the system 1.1 State, purpose, general statement AOT (Ontology Alignment at Terminological level) and AOTL (Ontology Alignment at Terminological and Linguistic level) are automatic ontology alignment systems realized for the purpose to solve the problem of ontology Matching. The AOT system uses various terminological matchers with a local filter in order to find correspondences between ontologies to be aligned. Contrary to AOT system, AOTL combines the similarities calculated by the various string matching algorithms at terminological level without a local filter, then these similarities are combined with similarities calculated using an external resource WordNet i.e. at linguistic level. The next step (for AOTL system) consists in combining the similarities by gives the priority to linguistic matcher; otherwise we have used an average aggregation method. Finally AOT and AOTL applied a filter in order to identify the alignment For AOT system, we have proposed a local filter (section 1.2.1.3) in order to select better correspondences and we envision to use AOT in order to study the system behavior using different aggregation and filter methods and proposing in the future more efficient filters and aggregation methods and add other matchers. For AOTL system, we have used an external resource WordNet in order to select semantic correspondences and we plan to use AOTL in order to discover new semantic correspondences more than select the better one i.e. we are interested in recall more than precision, of course with good the f-measure (balanced). The details of each step of our systems are described in the following section. 1.2 Specific techniques used The process of AOT and AOTL systems consists in the following two successive steps: 1) Calculation of Similarities and 2) Combination and Extraction of Alignment.