Shenghui Wang

An empirical study of instance-based ontology matching

Instance-based ontology mapping is a promising family of solutions to a class of ontology alignment problems. It crucially depends on measuring the similarity between sets of annotated instances. In this paper we study how the choice of co-occurrence measures affects the performance of instance-based mapping. To this end, we have implemented a number of different statistical co-occurrence measures. We have prepared an extensive test case using vocabularies of thousands of terms, millions of instances, and hundreds of thousands of co-annotated items. We have obtained a human Gold Standard judgement for part of the mapping-space. We then study how the different co-occurrence measures and a number of algorithmic variations perform on our benchmark dataset as compared against the Gold Standard. Our systematic study shows excellent results of instance-based matching in general, where the more simple measures often outperform more sophisticated statistical co-occurrence measures.

Learning Concept Mappings from Instance Similarity

Finding mappings between compatible ontologies is an important but difficult open problem. Instance-based methods for solving this problem have the advantage of focusing on the most active parts of the ontologies and reflect concept semantics as they are actually being used. However such methods have not at present been widely investigated in ontology mapping, compared to linguistic and structural techniques. Furthermore, previous instance-based mapping techniques were only applicable to cases where a substantial set of instances was available that was doubly annotated with both vocabularies. In this paper we approach the mapping problem as a classification problem based on the similarity between instances of concepts. This has the advantage that no doubly annotated instances are required, so that the method can be applied to any two corpora annotated with their own vocabularies. We evaluate the resulting classifiers on two real-world use cases, one with homogeneous and one with heterogeneous instances. The results illustrate the efficiency and generality of this method.

Concept drift and how to identify it

This paper studies concept drift over time. We first define the meaning of a concept in terms of intension, extension and label. Then we study concept drift over time using two theories: one based on concept identity and one based on concept morphing. A qualitative toolkit for analysing concept drift is proposed to detect concept shift and stability when concept identity is available, and concept split and strength of morphing chain if using the morphing theory. We apply our framework in four case-studies: a political vocabulary in SKOS, the DBpedia ontology in RDFS, the LKIF-Core ontology in OWL and a few biomedical ontologies in OBO. We describe ways of identifying interesting changes in the meaning of concept within given application contexts. These case-studies illustrate the feasibility of our framework in analysing concept drift in knowledge organisation schemas of varying expressiveness.

Matching multi-lingual subject vocabularies

Most libraries and other cultural heritage institutions use controlled knowledge organisation systems, such as thesauri, to describe their collections. Unfortunately, as most of these institutions use different such systems, unified access to heterogeneous collections is difficult. Things are even worse in an international context when concepts have labels in different languages. In order to overcome the multilingual interoperability problem between European Libraries, extensive work has been done to manually map concepts from different knowledge organisation systems, which is a tedious and expensive process. Within the TELplus project, we developed and evaluated methods to automatically discover these mappings, using different ontology matching techniques. In experiments on major French, English and German subject heading lists Rameau, LCSH and SWD, we show that we can automatically produce mappings of surprisingly good quality, even when using relatively naive translation and matching methods.

Evaluating Thesaurus Alignments for Semantic Interoperability in the Library Domain

Thesaurus alignments play an important role in realizing efficient access to heterogeneous cultural-heritage data. Current technology, however, provides only limited value for such access because it fails to bridge the gap between theoretical study and practical application requirements. This article explores common real-world library problems and identifies solutions that focus on the application-embedded study, development, and evaluation of matching technology.

Two variations on ontology alignment evaluation: methodological issues

Evaluation of ontology alignments is in practice done in two ways: (1) assessing individual correspondences and (2) comparing the alignment to a reference alignment. However, this type of evaluation does not guarantee that an application which uses the alignment will perform well. In this paper, we contribute to the current ontology alignment evaluation practices by proposing two alternative evaluation methods that take into account some characteristics of a usage scenario without doing a full-fledged end-to-end evaluation. We compare different evaluation approaches in three case studies, focussing on methodological issues. Each case study considers an alignment between a different pair of ontologies, ranging from rich and well-structured to small and poorly structured. This enables us to conclude on the use of different evaluation approaches in different settings.

Enhancing content-based recommendation with the task model of classification

In this paper, we define reusable inference steps for content-based recommender systems based on semantically-enriched collections. We show an instantiation in the case of recommending artworks and concepts based on a museum domain ontology and a user profile consisting of rated artworks and rated concepts. The recommendation task is split into four inference steps: realization, classification by concepts, classification by instances, and retrieval. Our approach is evaluated on real user rating data. We compare the results with the standard content-based recommendation strategy in terms of accuracy and discuss the added values of providing serendipitous recommendations and supporting more complete explanations for recommended items.

Deriving Concept Mappings through Instance Mappings

Ontology matching is a promising step towards the solution to the interoperability problem of the Semantic Web. Instance-based methods have the advantage of focusing on the most active parts of the ontologies and reflect concept semantics as they are actually being used. Previous instance-based mapping techniques were only applicable to cases where a substantial set of instances shared by both ontologies. In this paper, we propose to use a lexical search engine to map instances from different ontologies. By exchanging concept classification information between these mapped instances, an artificial set of common instances is built, on which existing instance-based methods can apply. Our experiment results demonstrate the effectiveness and applicability of this method in broad thesaurus mapping context.

What is concept drift and how to measure it

This paper studies concept drift over time. We first define the meaning of a concept in terms of intension, extension and label. We then introduce concept drift over time and two derived notions: (in)stability over a time period and concept shift between two time points. We apply our framework in three case-studies, one from communication science, on DBPedia, and one in the legal domain. We describe ways of identifying interesting changes in the meaning of concept within given application contexts. These case-studies illustrate the feasibility of our framework in analysing concept drift in knowledge organisation schemas of varying expressiveness.

Instance-Based Ontology Matching by Instance Enrichment

The ontology matching (OM) problem is an important barrier to achieve true Semantic Interoperability. Instance-based ontology matching (IBOM) uses the extension of concepts, the instances directly associated with a concept, to determine whether a pair of concepts is related or not. While IBOM has many strengths it requires instances that are associated with concepts of both ontologies, (i.e) dually annotated instances. In practice, however, instances are often associated with concepts of a single ontology only, rendering IBOM rarely applicable. In this paper we discuss a method that enables IBOM to be used on two disjoint datasets, thus making it far more generically applicable. This is achieved by enriching instances of each dataset with the conceptual annotations of the most similar instances from the other dataset, creating artificially dually annotated instances. We call this technique instance-based ontology matching by instance enrichment (IBOMbIE). We have applied the IBOMbIE algorithm in a real-life use-case where large datasets are used to match the ontologies of European libraries. Existing gold standards and dually annotated instances are used to test the impact and significance of several design choices of the IBOMbIE algorithm. Finally, we compare the IBOMbIE algorithm to other ontology matching algorithms.