1 / 65

Ontology Mapping

Ontology Mapping. Elham Paikari paikari@ce.sharif.edu Sharif University Of Technology Computer Engineering Department. Agenda. The Role of Ontology Ontology Integration About the Problem Ontology Mismatch Language Level Mismatches Ontology Level Mismatches Conceptualization Mismatch

fulton-carr
Download Presentation

Ontology Mapping

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Ontology Mapping Elham Paikari paikari@ce.sharif.edu Sharif University Of Technology Computer Engineering Department Semantic Web, Ontology Mapping Fall 2005

  2. Agenda • The Role of Ontology • Ontology Integration • About the Problem • Ontology Mismatch • Language Level Mismatches • Ontology Level Mismatches • Conceptualization Mismatch • Components Of Mapping • Similarity Calculation • Further Refinements • Path Length Measurement • Equivalence • Interoperability Semantic Web, Ontology Mapping Fall 2005

  3. The Role of Ontology • The word ontology comes from the Greek ontos for being. It is a relatively new term in the long history of philosophy, introduced by the 19th century German philosophers to distinguish the study of being. • In information systems, a more pragmatic view to ontology is taken, where ontology is considered as a kind of agreement on a domain representation: ontology is an explicit account or representation of a conceptualization. This conceptualization includes a set of concepts, their definitions and their inter-relationships. Semantic Web, Ontology Mapping Fall 2005

  4. Ontology Integration Many works on ontology comparison has been motivated by ontology integration: given a set of independently developed ontologies, construct a single global ontology. The first step in integrating the ontologies is: • Identify and characterize inter-ontology correspondences. Semantic Web, Ontology Mapping Fall 2005

  5. Ontology Integration • The starting point for comparing and mapping heterogeneous semantics in ontology mapping is to semantically enrich the ontologies. • Semantic enrichment facilitates ontology mapping by making explicit different kinds of ”hidden” information concerning the semantics of the modeled objects. The underlying assumption is that the more semantics that are explicitly specified about the ontologies, the more feasible their comparison becomes. Semantic Web, Ontology Mapping Fall 2005

  6. Classification Of Ontology Specification Language Semantic Web, Ontology Mapping Fall 2005

  7. Life Cycle Of An Ontology Semantic Web, Ontology Mapping Fall 2005

  8. A Generic Architecture Of Ontology-Based Applications Semantic Web, Ontology Mapping Fall 2005

  9. Beneficial Applications • Semantic Web • Knowledge Management • Information Retrieval • Service Retrieval Semantic Web, Ontology Mapping Fall 2005

  10. About the Problem The Semantic Web proposes to standardize a semantic markup method for: • Uniform formalism, XML • Organization of knowledge into ontologies The scientific difficulties are linked to • Exact definition of the formalisms • Impossibility of maintaining a worldwide centralization of the ontologies Other challenges concern • Robustness • Scalability of these techniques Semantic Web, Ontology Mapping Fall 2005

  11. Survey We start with an introduction to the problem of ontology heterogeneity, which is characterized by different kinds of mismatches between ontologies. This kind of heterogeneity hampered us from a combined usage of multiple ontologies, which is needed in many applications. To solve the heterogeneity problem, the mismatches need to be reconciled. This means that we need to map and align different ontologies. Semantic Web, Ontology Mapping Fall 2005

  12. Ontology Mismatch Differences between ontologies are called mismatches, Issues: • Practical problems • Ontologies • Versioning The main concern here is mismatches between ontologies. They are further divided into language level and ontology level. The former conforms to the syntactic layer, and the latter to the semantic layer. Semantic Web, Ontology Mapping Fall 2005

  13. Ontology Heterogeneity Semantic Web, Ontology Mapping Fall 2005

  14. Language Level Mismatches Mismatches at the language level occur when ontologies written in different ontology languages are combined. Four types of mismatches are identified. • Syntax. Different ontology languages often use different syntaxes. For example, to define the class of car in RDF Schema, one uses <rdfs:Class ID = "Car">. In LOOM, the expression (defconcept Car) is used to define the same class. • Logical representation. A slightly more complicated mismatch at this level is the difference in representation of logic notions. For example, in some languages, it is possible to state explicitly that two classes are disjoint (e.g. disjoint A B), whereas it is necessary to use negation in subclass statements in other languages (e.g. A subclassof (NOT B), B subclass-of (Not A)) Semantic Web, Ontology Mapping Fall 2005

  15. Language Level Mismatches • Semantics of primitives. A more subtle possible difference at the language level is the semantics of language constructs. Despite the fact that sometimes the same name is used for a language construct in two languages, the semantics may differ, e.g., there are several interpretation of A equalTo B. • Language expressivity. The mismatch at the language level with the most impact is the difference in expressivity between two languages. This difference implies that some languages are able to express things that are not expressible in other languages. For example, some languages have constructs to negation, whereas others have not. Semantic Web, Ontology Mapping Fall 2005

  16. Ontology Level Mismatches Mismatches at the ontology level happen when two or more ontologies that describe partly overlapping domains are combined. In the same language, or different languages. • conceptualization mismatch is a difference in the way a domain is interpreted. divided into model coverage and concept scope (granularity). • Explication mismatch is a difference in the way the conceptualization is specified. Semantic Web, Ontology Mapping Fall 2005

  17. Conceptualization Mismatch • Scope. Two classes seem to represent the same concept, but do not have the same instances, although they may intersect. The classical example is the class ”employee”, where several administrations use slightly different concepts of employee. • Model coverage and granularity. This is a mismatch in the part of the domain that is covered by the ontology, or the level of detail to which that domain is modeled. the example of an ontology about cars: one ontology might model cars but not trucks. Another one might represent trucks but only classify them into a few categories, while a third ontology might make very finegrained distinctions between types of trucks based on their physical structure, weight, purpose, etc. Semantic Web, Ontology Mapping Fall 2005

  18. Explication Mismatches • Two types of differences can be classified as terminological mismatches. – Synonym terms. Concepts are represented by different names. One example is the use of term ”car” in one ontology and the term ”automobile” in another ontology. – Homonym terms. The meaning of the same term is different in different context. For example, the term ”conductor” has a different meaning in a music domain than it has in an electric engineering domain. Semantic Web, Ontology Mapping Fall 2005

  19. Explication Mismatches • Modeling style is related to the paradigm and conventions taken by the developers. – Paradigm. Different paradigms can be used to represent concepts such as time, action, plans, causality, propositional attitudes, etc. For example, one model might use temporal representations based on interval logic while another might use a representation based on point. – Concept description. This type of differences are called modeling conventions. Several choices can be made for the modeling of concepts in the ontologies. For example, a distinction between two classes can be modeled using a qualifying attribute or by introducing separate class. • Encoding mismatches are differences in value formats, like measuring distance in miles or in kilometers. Semantic Web, Ontology Mapping Fall 2005

  20. Approaches and Techniques • The focus of this work is on ontology level mismatch (semantic mismatch). • There are also approaches to tackle syntactic mismatches. We will briefly describe some of those in order to give a complete picture of the state. Semantic Web, Ontology Mapping Fall 2005

  21. Solving Language Mismatches Four approaches to enable interoperability between different ontologies at the language level have been identified. • Aligning the metamodel. The constructs in the language are formally specified in a general model. • Layered interoperability. Aspects of the language are split up in clearly defined layers, and interoperability is to be resolved layer by layer. Semantic Web, Ontology Mapping Fall 2005

  22. Solving Language Mismatches • Transformation rules. The relation between two specific constructs in different ontology language is described in the form of a rule that specifies the transformation from the one to the other. • Mapping onto a common knowledge model. The constructs of an ontology language are mapped onto a common knowledge model, e.g. OKBC (Open Knowledge Base Connectivity). Semantic Web, Ontology Mapping Fall 2005

  23. Solving Ontology Level Mismatches The alignment of concepts at the ontology level requires • Understanding of the meaning of concepts • Cannot be fully automated. At the model level, there exist mainly tools that suggest alignments and mappings based on heuristics matching algorithm and provide means to specify these mappings. Semantic Web, Ontology Mapping Fall 2005

  24. Solving Ontology Level Mismatches Finally, in order to integrate ontologies, it is important to distinguish mismatches that are hard to solve, and those that are not. conceptualization mismatches often need human intervention to be solved. Most explication mismatches can be solved automatically, but the terminological mismatches may be difficult. Encoding mismatches can be quite easily solved with a transformation step. Semantic Web, Ontology Mapping Fall 2005

  25. Definition And Scope Of Ontology Mapping The concept of ”mapping” has a range of meanings, including integration, unification, merging, etc. • Mapping will be a set of formulate that provide the semantic relationships between the concepts in the models. • Mapping is to establish correspondences among the source ontologies, and to determine the set of overlapping concepts, concepts that are similar in meaning but have different names or structure, and concepts that are unique to each of the sources. Semantic Web, Ontology Mapping Fall 2005

  26. Definition And Scope Of Ontology Mapping • Merging and is to create a single coherent ontology that includes the information from all the sources. • Alignment is to make the source ontologies consistent and coherent with one another but kept separately. The aim of mapping is to map concepts in the various ontologies to each other, so that a concept in one ontology corresponds to a query (i.e. view) over the other ontologies Semantic Web, Ontology Mapping Fall 2005

  27. Definition And Scope Of Ontology Mapping Two tasks have to be conducted in the ontology mapping process: • Discover the correspondences between ontology elements (1) Applying a set of matching rules (2) Evaluating interesting similarity measures that compare a set of possible correspondence and help to choose valid correspondence from them. • Describe and define the discovered mappings so that other follow-up components could make use of them Semantic Web, Ontology Mapping Fall 2005

  28. Application Domains • Information Integration and the Semantic Web. In many contexts, data resides in a multitude of data sources. In the Semantic Web context, an ontology captures the semantics of data. Data integration enables users to ask queries in a uniform fashion, without having to access each data source independently. In addition to query, mappings between ontologies are necessary for agents to interoperate. Semantic Web, Ontology Mapping Fall 2005

  29. Application Domains • Ontology merging. Several applications require that we combine multiple ontologies into a single coherent ontology. In some cases, these are independently developed ontologies that model overlapping domains. In others, we merge two ontologies that evolved from a single base ontology. Semantic Web, Ontology Mapping Fall 2005

  30. Automatic Ontology Mapping Tools • Automated tools can significantly speed up the process by proposing plausible mappings • Some parts need expert intervention Approaches for building such tools • Use a wide range of heuristics to generate mappings • Learn mappings Semantic Web, Ontology Mapping Fall 2005

  31. Systems for Ontology Mapping First present Chimaera, a webbased ontology merging and diagnosing environment. Then, present PROMPT, an algorithm used in Prot´eg´e for ontology merging. Next is FCA-Merge, which merges ontologies using documents on the same domain for the ontologies to be merged. MOMIS, which merges ontologies by means of ontology clustering finally we present GLUE, which performs ontology mapping by machine learning techniques. Semantic Web, Ontology Mapping Fall 2005

  32. Characteristics of ontology mapping systems Semantic Web, Ontology Mapping Fall 2005

  33. Characteristics of ontology mapping systems Semantic Web, Ontology Mapping Fall 2005

  34. Algorithm Overview The basic idea of mapping assertion analysis applied in practice for comparison of relevant elements of two ontologies. Semantic Web, Ontology Mapping Fall 2005

  35. Components Of Mapping • The Mapper performs a computation of a correspondence measure for the pairs of compared ontology elements, based on the similarity of their enriched structures. • The Enhancer utilizes an electronic lexicon to adjust the similarity values that have been computed by the mapper, with the intention of re-ranking the mapping assertions in the result list. • The Presenter determines which recommendation to suggest to the user, based on the partial ordering of correspondence measures. Semantic Web, Ontology Mapping Fall 2005

  36. Components Of Mapping • The Exporter translates and exports the mapping results to a desired format so that other follow-up applications can import and use the results in a loosely coupled way. • The Configuration Profile is a user profile to assign individual variable values for different tuning parameters and a threshold value for exclusion of mappings with low similarity. Semantic Web, Ontology Mapping Fall 2005

  37. Mapping Algorithm The mapping algorithm is used to: • Semi-Automate the process • Comparing & mapping two semantically enriched ontologies • The algorithm produces a set of ranked suggestions. The user is in control of accepting, rejecting or altering the assertions. The level of automatic exclusion from user presentation is adjustable. Semantic Web, Ontology Mapping Fall 2005

  38. Similarity Calculation for Concepts The similarity of two concepts in two ontologies is directly calculated as the cosine measure between the two representative feature vectors. Let two feature vectors for concept a and b respectively, both of length n, be given. The cosine similarity between concept a and concept b is defined as: Semantic Web, Ontology Mapping Fall 2005

  39. Similarity Calculation for Concepts Ca and Cb are feature vectors for concept a and b, respectively • n is the dimension of the feature vectors • |Ca| and |Cb| are the lengths of the two vectors A threshold value is defined by the user to exclude pairs that have too low similarity values. Semantic Web, Ontology Mapping Fall 2005

  40. Path Length Measurement • Path length is measured in nodes rather than links • in WordNet, nouns are organized into taxonomies where each node is a set of synonyms (a synset) representing a single sense. • If a word has multiple senses, it will appear in multiple synsets at various locations in the taxonomy. • Verbs are structured in a similar hierarchy with the relation being troponymy in stead of hypernymy. Semantic Web, Ontology Mapping Fall 2005

  41. Path Length Measurement We use in this experiment is that of hyponymy/hypernymy, or the is-a-kind-of relation, which relates more general and more specific senses. One way to measure the semantic similarity between two words a and b is to measure the distance between them in WordNet. This can be done by finding the paths from each sense of a to each sense of b and then selecting the shortest such path. Semantic Web, Ontology Mapping Fall 2005

  42. Example on hyponymy relation in WordNet used for the path length measurement Semantic Web, Ontology Mapping Fall 2005

  43. Similarity Calculation for Complex Elements Based on the correspondences calculated for the concepts, we could further expand the correspondence discovery into other elements and structures in the ontologies. Semantic Web, Ontology Mapping Fall 2005

  44. Relations The similarity of relations is calculated based on the corresponding domain concepts and range concepts of the relations. • X and X’ are domain concepts of R and R’ • Y and Y’ are the range concepts of R and R’ • the sim(X, X0) and sim(Y, Y0) can be calculated by equation for concepts similarity. Semantic Web, Ontology Mapping Fall 2005

  45. Clusters For this, we define the concept of cluster. • A cluster is a group of related concepts, which includes a center concept a and its k-nearest neighbors. • A cluster of 1-nearest neighbor includes a center concept and its direct parent, and its direct children. • A cluster of 2-nearest neighbor includes the grandparent, the siblings and the grandchildren, in addition to the 1-nearest neighbor. Semantic Web, Ontology Mapping Fall 2005

  46. Example of calculating cluster similarity Semantic Web, Ontology Mapping Fall 2005

  47. Similarity Of Clusters The similarity of clusters is calculated based on the weighted percentage of established mappings between member concepts in proportion to the number of all connections between the two clusters. The similarity between cluster A and cluster B is therefore computed as: Semantic Web, Ontology Mapping Fall 2005

  48. Similarity Of Clusters • X and Y are clusters of k-nearest neighbor. X = {a1, a2, a3, · · · , an} and Y = {b1, b2, b3, · · · , bm}. • M is a subset of the cartesian product of X and Y, where M ⊆ X * Y,M = {(ai, bj)|(ai ∊ X) ∩ (bj ∊ Y) ∩ (sim(ai, bj) > 0)} • |X| and |Y| are number of elements in the two sets, respectively. • The sim(ai, bj) is calculated by equation for concepts similarity. Semantic Web, Ontology Mapping Fall 2005

  49. Ontologies The similarity between two ontologies can be quantified as the weighted percentage of established mappings in proportion to all the connections between concepts in the two ontologies, as defined in the following equation. Semantic Web, Ontology Mapping Fall 2005

  50. Similarity Of Ontologies • O1 and O2 are two ontologies. O1 = {a1, a2, a3, · · · , an} and O2 = {b1, b2, b3, · · · , bm}. ai (i=1...n) are the concepts in O1 and bj (j=1...m) are the concepts in O2. • M is a subset of the cartesian product of O1 and O2, where M ⊆ O1 * O2,M = {(ai, bj)|(ai ∊ O1) ∩ (bj ∊ O2) ∩ (sim(ai, bj) > 0)} • |O1| and |O2| are number of concepts in the two ontologies, respectively. • The sim(ai, bj) is calculated by equation 6.1 for concepts similarity. Semantic Web, Ontology Mapping Fall 2005

More Related