Using AgreementMaker to align ontologies for OAEI 2010

The AgreementMaker system is unique in that it features a powerful user interface, a flexible and extensible architecture, an integrated evaluation engine that relies on inherent quality measures, and semi-automatic and automatic methods. This paper describes the participation of AgreementMaker in the 2010 OAEI competition in three tracks: benchmarks, anatomy, and conference. After its successful participation in 2009, where it ranked first in the conference track, second in the anatomy track, and obtained good results in the benchmarks track, the goal in this year’s participation is to increase the values of precision, recall, and F-measure for each of those tracks. 1 Presentation of the system We have been developing the AgreementMaker system since 2001, with a focus on real-world applications [5, 8] and in particular on geospatial applications [4, 6, 7, 9–13]. However, the current version of AgreementMaker, whose development started two years ago, represents a whole new effort. 1.1 State, purpose, general statement The new AgreementMaker system [1–3] supports: (1) user requirements, as expressed by domain experts; (2) a wide range of input (ontology) and output (agreement file) formats; (3) a large choice of matching methods depending, on the different granularity of the set of components being matched (local vs. global), on different features considered in the comparison (conceptual vs. structural), on the amount of intervention that they require from users (manual vs. automatic), on usage (standalone vs. composed), and on the types of components to consider (schema only or schema and instances); (4) improved performance, that is, accuracy (precision, recall, F-measure) and efficiency (execution time) for the automatic methods; (5) an extensible architecture to incorporate new methods easily and to tune their performance; (6) the capability to evaluate, compare, and combine different strategies and matching results; (7) a comprehensive user interface that supports advanced visualization techniques and a control panel that ? Research supported by NSF Awards IIS-0513553 and IIS-0812258. ?? Additional affiliation: University of Milan-Bicocca, Italy. drives all the matching methods and evaluation strategies; (8) a feedback loop that accepts suggestions and corrections by users and extrapolates new mappings. In 2009 AgreementMaker was very successful in the OAEI competition. In particular, AgreementMaker ranked (a close) second among ten systems in the anatomy track. AgreementMaker also participated successfully in two other tracks: benchmarks and conference. In the former track, AgreementMaker was ranked first in terms of precision and seventh in terms of recall among thirteen systems and in the latter track AgreementMaker was ranked first with the highest F-measure (57% at a threshold of 75%) among seven competing systems. 1.2 Specific techniques used AgreementMaker comprises several matching algorithms or matchers that can be used for matching (or aligning) the source and target ontologies. The matchers are not restricted to any particular domain. The architecture of AgreementMaker relies on a stack of matchers that belong to three different layers (see Figure 1). Specific configurations of the stack have been used for the benchmarks, anatomy, and conference tracks, as discussed in what follows. However, we describe first the different components in the stack: the matchers, the combination and evaluation modules, and the final alignment module. Fig. 1. AgreementMaker OAEI 2010 matcher stack. Matchers can be concept-based (if they consider only one concept) or structural (if they consider a subgraph of the ontology). The concept-based matchers support the comparison of strings. They include: the Base Similarity Matcher (BSM) [7], the Parametric String-based Matcher (PSM) [2] and the Vector-based Multi-Word Matcher (VMM) [2]. BSM is a basic string matcher that computes the similarity between concepts by comparing all the strings associated with them. PSM is a more in-depth string matcher, which for the competition is set to use a substring measure and an edit distance measure. VMM compiles a virtual document for every concept of an ontology, transforms the resulting strings into TF-IDF vectors and then computes their similarity using the cosine similarity measure. These matchers have been extended in the AgreementMaker configuration used this year by plugging in a set of lexicons, which are used to expand the set of strings with synonyms. The extended matchers are therefore called BSM, PSM, and VMM. The Advanced Similarity Matcher (ASM) is a string-based matcher that computes mappings between source and target concepts (including their properties) by comparing their local names, and providing better similarity evaluation in particular when compound terms are used. ASM outperforms generic string-based similarity matchers because it is based on a deeper linguistic analysis. Structural matchers include the Descendants’ Similarity Inheritance (DSI) matcher [7]. This matcher is based on the idea that if two nodes are similar, then their descendants should be similar. The Group Finder Matcher (GFM) is another structural matcher that filters out the mappings provided by another matcher (the input matcher). It identifies groups of concepts and properties in the ontologies and assumes that two concepts (or properties) that belong to two groups that were not mapped by the input matcher will likely have different meanings and should not be mapped. The Iterative Instance Structural Matcher (IISM) takes into account instances. Classes that have mapped individuals can then be aligned. In addition, values of the properties are also considered. The structural part of IISM is quite complex and takes into account superclasses, subclasses, properties, subproperties, cardinalities, and the range and domain of properties. The combination and evaluation modules are used together, as follows. The Linear Weighted Combination (LWC) [2] combines its inputs (e.g., from several string matchers), using a local confidence quality measure provided by the evaluation module, in order to automatically assign weights to each result computed by the input matchers. After this step, we have a single combined set of alignments that includes the best alignments from each of the input matchers. The final alignment module is given as input a mapping cardinality (e.g., 1:1) and a threshold and outputs the best set of alignments given those two inputs [2]. Benchmarks For the benchmarks track we used the following configuration: IISM( LWC(ASM,PSM,VMM,BSM) ) LWC is adopted to combine the results of four string-based matchers, namely ASM, PSM, PSM, and BSM; the last three make use of two lexicons, namely WordNet and a dictionary built from the ontologies; the similarity values computed at this step are then given as input to the IISM structural matcher. Anatomy For the anatomy track we used the following configuration: LWC(PSM,VMM,BSM) LWC is adopted to combine the results of four string-based matchers, namely PSM, VMM, and BSM; the last three make use of two lexicons, namely WordNet and a dictionary built from the ontologies. Conference For the conference track we used the following configuration: