Ontology Learning for the Semantic Web

OntoCase is a framework for semi-automatic pattern-based ontology construction. In this paper we focus on the retain and reuse phases, where an initial ontology is enriched based on content ontology design patterns (Content ODPs), and especially the implementation and evaluation of these phases. Applying Content ODPs within semi-automatic ontology construction, i.e. ontology learning (OL), is a novel approach. The main contributions of this paper are the methods for pattern ranking, selection, and integration, and the subsequent evaluation showing the characteristics of ontologies constructed automatically based on ODPs. We show that it is possible to improve the results of existing OL methods by selecting and reusing Content ODPs. OntoCase is able to introduce a general top structure into the ontologies, and by exploiting background knowledge the ontology is given a richer overall structure.KeywordsString MatchText CorpusOntology LearningOntology ConstructionOntology EngineerThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

In this paper we explore some of the opportunities and challenges for machine learning on the Semantic Web. The Semantic Web provides standardized formats for the representation of both data and ontological background knowledge. Semantic Web standards are used to describe meta data but also have great potential as a general data format for data communication and data integration. Within a broad range of possible applications machine learning will play an increasingly important role: Machine learning solutions have been developed to support the management of ontologies, for the semi-automatic annotation of unstructured data, and to integrate semantic information into web mining. Machine learning will increasingly be employed to analyze distributed data sources described in Semantic Web formats and to support approximate Semantic Web reasoning and querying. In this paper we discuss existing and future applications of machine learning on the Semantic Web with a strong focus on learning algorithms that are suitable for the relational character of the Semantic Web's data structure. We discuss some of the particular aspects of learning that we expect will be of relevance for the Semantic Web such as scalability, missing and contradicting data, and the potential to integrate ontological background knowledge. In addition we review some of the work on the learning of ontologies and on the population of ontologies, mostly in the context of textual data.

As semantic web technologies are increasingly used to empower knowledge management systems (KMSs), there is a growing need for mechanisms and automated tools for checking content generated by semantic-web tools. The content in a KMS includes both the knowledge management (KM) schema and the data contained within. KM schemas can be viewed as ontologies and the data contained within can be viewed as instance data. Thus we can apply semantic web ontology and instance data processing techniques and tools in KM settings. There are many semantic web tools aimed at ontology evaluation, however there is little, if any, research focusing on instance data evaluation. Although instance data evaluation has many issues in common with ontology evaluation, there are some issues that are either more prominent in or unique to instance data evaluation. Instance data often accounts for orders of magnitude more data than ontology data in organization intranets, thus our work focuses on evaluation techniques that help users of KMSs to determine when certain instance data is ready for use. We present our work on semantic web instance data evaluation for KMSs. We define the instance data evaluation research problem and design a general evaluation process GEP. We identify three categories of issues that may occur in instance data: syntax errors, logical inconsistencies, and potential issues. For each category of issues, we provide illustrative examples, describe the symptoms, analyze the causes, and present our detection solution. We implement our design in TW OIE which is an online instance data evaluation service. We perform experiments that show that the TW OIE is more comprehensive than most existing online semantic web data evaluators.

Trying out or updating an existing learning ontology and providing evaluation tools to assess its quality are essential steps before putting an e-learning system online. Ontology evaluation is a crucial task and it is usually the output of an automatic process. This paper proposes an ontology-driven model, called Ontology-Driven model for the Evaluation of Learning Ontologies (ODELO), for the evaluation of ontologies representing learning resources with respect to syntactic, semantic, pragmatic, social and cohesion metrics. Syntax deals with the formal relations between signs (e.g., words, phrases, sentences) and the production of new ones. Semantics is the study of the relationships between the system of signs and their meanings. Pragmatics is the study of natural language understanding. Social metrics reflect the fact that software agents and ontologies coexist and communicate in communities. Ontology cohesion metrics refer to the degree of relatedness of ontology classes. ODELO is a deductive evaluation model that identifies the elements of ontological quality for learning ontologies. In this paper we propose a framework for assessing the quality of learning ontologies, which constitute the basis for intelligent educational Adaptive Hypermedia (AH) systems. We introduce a new pragmatic quality metric for assessing the quality of learning ontologies.

In this paper we present a new method based on semi-automatic ontology construction that can be used to improve the understandability of legal documents. Legal documents typically extensively define requirements and procedures in a specific legislative area; usually, they are hard to comprehend for citizens without a proper legal knowledge. However, a vast majority of today’s e-Government activities for citizens (G2C) are governed by legal documents. Therefore, by improving the citizen’s comprehension many intricacies that occasionally occur during G2C activities can be avoided. Our method first divides a legal document into several paragraphs. From the paragraphs it semi-automatically constructs ontology of the field by using a tool OntoGen. Ontology concepts are then used to classify each paragraph and the resulting classification is visualized in a simple matrix, where rows represent paragraphs and columns represent top-level ontology concepts. Based on the visualization, paragraphs that need revising are identified; they can be relocated to more suitable context within the document or rewritten using more appropriate wording. We demonstrate the presented method on the document defining the tender for selling flats at favorable prices at the Housing Fund of the Republic of Slovenia, a public fund. We argue that by using the new method we were able to substantially improve the comprehension of the document. In addition, the constructed ontology helped the Fund’s officers improve the structure of their knowledge about the underlying business process.Keywordsontology constructionknowledge managementlegal documents

The Semantic Web offers significant capabilities that transform the current Web into a global knowledge base including various cross-linked multimedia content with formal descriptions of its semantics understandable to humans and processable by computers. Content on the Semantic Web can be subject to reasoning and queries with standardized languages, methods and tools, which opens new opportunities for collaborative creation, use and exploration of web repositories. However, these opportunities have not been exploited so far by the available 3D formats and modeling tools, which limits the possibilities of search and reuse of 3D content as part of the Semantic Web. This work contributes a semantic development pipeline of the X3D Ontology, with corresponding conversion of X3D models into triple forms suitable for formal query. The ontology design reflects experience accompanying the development of the Extensible 3D (X3D) Graphics International Standard, in particular, the X3D Unified Object Model (X3DUOM). This approach combines semantic and syntactic elements of X3D models and metadata to support integration with the Semantic Web. The pipeline enables automatic generation of the X3D Ontology, thereby providing an up-to-date 3D representation with semantics during X3D specification development. By extending commonplace model conversions from other formats to X3D, the ontology presents the potential to enable integration of most forms of 3D content with the Semantic Web.

Ontologies and knowledge bases are fundamental cornerstones in semantic search systems based on which efficient web services and sophisticated search mechanisms are designed. High quality domain ontologies are essential for successful employment of semantic Web services. However, acquisition of ontologies is difficult and costly, thus hindering the development of this field. In this work, the knowledge acquisition problem in the semantic Web is cast as the task of ontology construction. We construct an ontology for “polycystic kidney disease” from domain text by a pattern based method which requires minimum human intervention. We also create ontology manually to present the evaluation results of the constructed ontology.KeywordsOntologyontology constructionpatternsconcept extraction

Ontology evaluation is required before using the ontology within applications. Similar with software practice, the purpose of ontology evaluation is to identify the achievement of requirement criteria. Users who require coverage criteria often seeking ontology that contain the terms related to their focused domain knowledge. Users encounter the difficulty to select a suitable ontology from variety of ontology evaluation approaches. Conceptualization of information related to ontology evaluation helps to identify the important component within ontology that helps towards coverage criteria achievement. This work proposes an algorithm to extract ontology documents gained from public ontology repositories like Falcons into its vocabulary parts focused on classes and literals. The algorithm then processes the extracted ontology components with similarity algorithm and later displays the result on the coverage match of ontology with provided terms and the terms that are synonym expanded using WordNet.

Most of the challenges faced when building the Semantic Web require a substantial amount of human labor and intelligence. Despite significant advancement in ontology learning and human language technology, the tasks of ontology construction, semantic annotation, and establishing alignments between multiple ontologies remain highly dependent on human intelligence. This means that individuals need to contribute time and sometimes other resources. Unfortunately, we observe a serious lack of user involvement in the aforementioned tasks, which may be due to the absence of motivations for people who contribute. As a novel solution, we (1) propose to masquerade the core tasks of weaving the Semantic Web behind online, multi-player game scenarios, in order to create proper incentives for human users to get involved. Doing so, we adopt the findings from the already famous “games with a purpose” by von Ahn, who has shown that presenting a useful task, which requires human intelligence, in the form of an online game can motivate a large amount of people to work heavily on this task, and this for free. Then, we (2) describe our generic OntoGame platform, and (3) several gaming scenarios for various tasks plus our respective prototypes. Based on the analysis of user data and interviews with players, we provide preliminary evidence that users (4) enjoy the games and are willing to dedicate their time to those games, (5) are able to produce high-quality conceptual choices. Eventually we show how users entertaining themselves by online games can unknowingly help weave and maintain the Semantic Web.

The explosive growth of textual data on the web coupled with the increase on demand for ontologies to promote the semantic web, have made the automatic ontology construction from the text a very promising research area. Ontology learning (OL) from text is a process that aims to (semi-) automatically extract and represent the knowledge from text in machine-readable form. Ontology is considered one of the main cornerstones of representing the knowledge in a more meaningful way on the semantic web. Usage of ontologies has proven to be beneficial and efficient in different applications (e.g. information retrieval, information extraction, and question answering). Nevertheless, manually construction of ontologies is time-consuming as well extremely laborious and costly process. In recent years, many approaches and systems that try to automate the construction of ontologies have been developed. This paper reviews various approaches, systems, and challenges of automatic ontology construction from the text. In addition, it also discusses ways the ontology construction process could be enhanced in the future by presenting techniques from shallow learning to deep learning (DL).

This work introduces a formal framework for the social acquisition of ontologies which are constructed dynamically from overhearing the possibly conflicting symbolic interaction of autonomous information sources, and an approach to the pragmatics of communicated ontological axioms. Technically, the framework is based on distributed variants of description logic for the formal contextualization of statements w.r.t. their respective provenance, speaker's attitude, addressees, and subjective degree of confidence. Doing so, our approach demarcates from the dominating more or less informal approaches to context and provenance representation on the semantic web, and carefully distinguishes between communication attitudes such as public assertion and intention exhibited on the (semantic) web on the one hand, and mental attitudes such as private belief on the other. Furthermore, our framework provides formal means for the probabilistic fusion of controversial opinions, and presents a semantics of information publishing acts. Our approach provides an incomplex and genuinely social approach to knowledge acquisition and representation, and is thus expected to be widely applicable in fields such as the semantic web and social software, and possibly also in other distributed environments such as P2P systems.

Ontology provides a shared and reusable piece of knowledge about a specific domain, and has been applied in many fields, such as semantic Web, e-commerce and information retrieval, etc. However, building ontology by hand is a very hard and error-prone task. Learning ontology from existing resources is a good solution. Because relational database is widely used for storing data and OWL is the latest standard recommended by W3C, this paper proposes an approach of learning OWL ontology from data in relational database. Compared with existing methods, the approach can acquire ontology from relational database automatically by using a group of learning rules instead of using a middle model. In addition, it can obtain OWL ontology, including the classes, properties, properties characteristics, cardinality and instances, while none of existing methods can acquire all of them. The proposed learning rules have been proven to be correct by practice.

The emergence of Semantic Web (SW) and the related technologies promise to make the Web a meaningful experience. Conversely, success of SW and its applications depends largely on utilization and interoperability of well-formulated ontology bases in an automated heterogeneous environment. This creates a need to investigate utilization of an (materialized) ontology view as an alternative version of an ontology. However, high level modeling, design and querying techniques still proves to be a challenging task for SW paradigm, as, unlike classical database systems, ontology view definitions and querying have to be done at high-level abstraction. In order to address such an issue, in this chapter, we describe an abstract view formalism forSW (SW-view) with conceptual and logical extensions. SW-views provides the needed conceptual and logical semantics to engineer ontology bases using three levels of abstraction, namely (1) conceptual, (2) logical/schema and (3) instance levels. We first provide the view model and it formal properties including a set of conceptual operators, which enable us to do ontology extraction at the conceptual level. Later, we provide a schemata transformation methodology to materialize SW-views under the Ontology Extraction Methodology (OEM) framework.

Semantic similarity is a metric used to know the similarity degree of two concepts in an ontology or a taxonomy. Semantic similarity has a wide variety applications on artificial intelligence, natural language processing, biomedical informatics, geoinformatics and semantic web and is usually applied on machine translation and word-sense disambiguation. In this research, semantic similarity measures are used to evaluate taxonomic relationships in a domain ontology. This evaluation was carried out by using a proposed algorithm and through the accuracy measure. The semantic similarity measures implemented are based on information content and were proposed by the following authors: Resnik, Lin, Jiang & Conrath and Mazandu & Mulder. Mainly, this research contributes to the automatic evaluation of ontologies in the task of evaluating the ontology taxonomy. The experimental results show that the measures have at least 88% accuracy. In addition, the system has an accuracy of 94% compared to validation responses from an expert.

Ontology is a representation term used to describe and represent a domain of knowledge. Manually ontology development is currently considered complex, requiring a lot of time and effort. This research was proposed to develop methods to build automatic domain ontology bilingual in Indonesian and English by using corpus and ontology design patterns (ODPs) in tuberculosis disease. In this study, the methods used were to combine ontology learning from text and correspond with ontology design patterns to decrease the role of expert knowledge. The methods in this research consist of six stages: (i) Term and relation extraction (ii) Corresponding with Tuberculosis glossary (iii) Corresponding the ontology design patterns (iv) Score computation similarity term and relations with ODPs (v) Ontology Building (vi) Ontology evaluation. The results of ontology construction were 361 terms and 44 relations with 260 terms were added. The calculation accuracy of ontology construction was 71%. Ontology construction had higher complexity and shorter time as well as decreases the role of the expert knowledge which proof that the automatic ontology evaluation is better than manual ontology construction.