Analysis of an Iterative Algorithm for Term-Based Ontology Alignment

Analysis of an Iterative Algorithm for Term-Based Ontology Alignment Shisanu Tongchim, Canasai Kruengkrai, Virach Sornlertlamvanich, Prapass Srichaivattana, and Hitoshi Isahara Thai Computational Linguistics Laboratory, National Institute of Information and Communications Technology, 112,Paholyothin Road, Klong 1, Klong Luang, Pathumthani 12120, Thailand {shisanu, canasai, virach, prapass}@tcllab.org, isahara@nict.go.jp Abstract. This paper analyzes the results of automatic concept align- ment between two ontologies. We use an iterative algorithm to perform concept alignment. The algorithm uses the similarity of shared terms in order to find the most appropriate target concept for a particular source concept. The results show that the proposed algorithm not only finds the relation between the target concepts and the source concepts, but the algorithm also shows some flaws in the ontologies. These results can be used to improve the correctness of the ontologies. 1 Introduction To date, several linguistic ontologies in different languages have been developed independently. The integration of these existing ontologies is useful for many applications. Aligning concepts between ontologies is often done by humans, which is an expensive and time-consuming process. This motivates us to find an automatic method to perform such task. However, the hierarchical structures of ontologies are quite different. The structural inconsistency is a common problem [1]. Developing a practical algorithm that is able to deal with this problem is a challenging issue. The objective of this research is to investigate an automated technique for ontology alignment. The proposed algorithm links concepts between two ontolo- gies, namely the MMT semantic hierarchy and the EDR concept dictionary. The algorithm finds the most appropriate target concept for a given source concept in the top-down manner. The experimental results show that the algorithm can find reasonable concept mapping between these ontologies. Moreover, the results also suggest that this algorithm is able to detect flaws and inconsistency in the ontologies. These results can be used for developing and improving the ontologies by lexicographers. The rest of this paper is organized as follows: Section 2 discusses related work. Section 3 provides the description of the proposed algorithm. Section 4 presents experimental results and discussion. Finally, Section 5 concludes our work. R. Dale et al. (Eds.): IJCNLP 2005, LNAI 3651, pp. 346–356, 2005. c ⃝ Springer-Verlag Berlin Heidelberg 2005 Analysis of an Iterative Algorithm for Term-Based Ontology Alignment 347 2 Related Work Daud´e et al. [2] used a relaxation labeling algorithm – a constraint satisfaction algorithm – to map the verbal, adjectival and adverbial parts between two dif- ferent WordNet versions, namely WordNet 1.5 and WordNet 1.6. The structural constraints are used by the algorithm to adjust the weights for the connections between WN1.5 and WN1.6. Later, some non-structural constraints are included in order to improve the performance [3]. Asanoma [4] presented an alignment technique between the noun part of WordNet and Goi-Taikei ’s Ontology. The proposed technique utilizes sets of Japanese and/or English words and semantic classes from dictionaries in an MT system, namely ALT-J/E. Chen and Fung [5] proposed an automatic technique to associate the English FrameNet lexical entries …

Content truncated to 500 words