Mapping Between Rdbms and Ontology a Review

Download Mapping Between Rdbms and Ontology a Review

Post on 01-Jun-2018

218 views

Category:

Documents

0 download

Embed Size (px)

TRANSCRIPT

<ul><li><p>8/9/2019 Mapping Between Rdbms and Ontology a Review</p><p> 1/7</p><p>INTERNATIONAL JOURNAL OF SCIENTIFIC &amp; TECHNOLOGY RESEARCH VOLUME 3, ISSUE 11, NOVEMBER 2014 ISSN 2277-8616 </p><p>307IJSTR2014www.ijstr.org</p><p>Mapping Between RDBMS And Ontology: AReview</p><p>Vishal Jain, Dr. S. V. A. V. Prasad</p><p>Abstract: Today Semantic web is playing a key role in the intelligent retrieval of information. It is the new-generation Web that tries to representinformation such that it can be used by machines not just for display purposes, but for automation, integration, and reuse across applications. It allowsthe representation and exchange of information in a meaningful way. Ontologies form the backbone of the Semantic Web; they allow machineunderstanding of information through the links between the information resources and the terms in the ontologies. Ontology describes basic concepts ina domain and defines relations among them. An ontology together with a set of individual instances of classes constitutes a knowledge base. An efforhas been made by the Semantic Web community to apply its semantic techniques in open, distributed and heterogeneous Web environments, and fosharing the knowledge in the semantic web. For sharing the knowledge ontologies were introduced, and have grown considerably in number. Buildingontology for a specific domain may be start from scratch or by modifying or using an existing ontology. The term Semantic Web (SW) given by TimBerners Lee is considered as vast concept within itself. Semantic Web (SW) is defined as collection of information linked in a way so that it can be easilyprocessed by machines. It is information in machine form. It contains Semantic Web Documents (SWDs) that are written in RD F or OWL languagesThey contain relevant information regarding users query. Crawlers play vital role in accessing information from SWDs..Index Terms: WWW, Semantic Web, Ontology, Ontology Mapping, OWL, RDBMS</p><p>1INTRODUCTION</p><p>Information Retrieval is the retrieval of information or data,either structured or unstructured. It retrieves in response toquery statement which may be unstructured or structured also.Unstructured Query is like sentence which is written incommon understandable language while structured query is inform of expression which is combination of equations andoperands. IR deals with fusion of streams of output documentsproduced from multiple retrieval methods. They combined toform single ranked stream which is shown to user. There aretwo methods for solving queries:</p><p>a) By submitting a given query to multiple documentcollections.</p><p>b) By submitting a given query through multiple IR methods.</p><p>Traditional text search engines fails for finding optimaldocuments because of following reasons:</p><p> Improper style of natural language: - These engines arenot capable of understanding complex way of writingdocuments.</p><p> High level unclear concepts: - Some concepts which areincluded in document but present search engines cantfind those words.</p><p> Semantic Relations: - We cant find relevant documents</p><p>for word specified in part of document. E.g. If we havesearched for Juice, then it will not find type or part ofJuice.</p><p>Information Retrieval mainly focuses on retrieval o</p><p>unstructured documents (natural text language documents)These documents may include videos, photos and audios etcIR addresses retrieval of documents from an organized weldefined huge collection of documents available on net whichmay be email, maps, news etc. Various goals of IR aredescribed below:</p><p> IR aims on retrieving unstructured documents. IR engine may produce collection of relevant documents</p><p>to user according to specified query entered by user. IR engine also arranges documents according to its rank</p><p>which involves Page Rank algorithm. If a document Ahas more effective results than document B, then A wilorganized first. It has been discussed in further sections.</p><p>2SEMANTIC WEBSemantic Web (SW) came into existence due to problem inconventional search engines that dissatisfies users byretrieving inadequate and inconsistent results. The documentsretrieved by conventional search engines are like horse odifferent colors. These engines work on predefined standardterms that work in centralized environment, thus accessingstandard Ontologies. With advent of SW and Ontology, usersare able to develop new facts and use their ownkeywords/terms in different environments. With use oontology, user can perform following tasks:</p><p>(a) Users can use Interface Description Languages (IDL</p><p>and services for different environments. IDL meansdefining new data objects and their relations.</p><p>(b) Users can communicate with different agents usingshared ontology like FOAF (Friend of a Friend).</p><p>Semantic Web (SW) is combination of SWDs that areexpressed in ontology languages (RDF, OWL). Ontology refersto categorization of concepts and relationships between termsin hierarchical fashion Although SWDs retrieves relevaninformation because they are characterized by semanticmethods and ideas, but it is tedious job to find URLs ofSWDs.</p><p>_______________________________</p><p> Vishal Jain, Research Scholar, Computer Science andEngineering Department, Lingayas University,</p><p>Faridabad, Haryana, INDIA.</p><p>E-mail:vishaljain83@ymail.com Dr. S. V. A. V. Prasad, Professor, Electronics and</p><p>Communication Department, Lingayas University,</p><p>Faridabad, Haryana, INDIA.E-mail:prasad.svav@gmail.com</p>mailto:vishaljain83@ymail.commailto:vishaljain83@ymail.commailto:vishaljain83@ymail.commailto:prasad.svav@gmail.commailto:prasad.svav@gmail.commailto:prasad.svav@gmail.commailto:prasad.svav@gmail.commailto:vishaljain83@ymail.com</li><li><p>8/9/2019 Mapping Between Rdbms and Ontology a Review</p><p> 2/7</p><p>INTERNATIONAL JOURNAL OF SCIENTIFIC &amp; TECHNOLOGY RESEARCH VOLUME 3, ISSUE 11, NOVEMBER 2014 ISSN 2277-8616 </p><p>308IJSTR2014www.ijstr.org</p><p>3WEBONTOLOGYLANGUAGE(OWL)W3Cs Web Ontology Working Group defined OWL as threedifferent sublanguages:</p><p>1. OWL Lite2. OWL DL (includes OWL Lite)3. OWL Full (includes OWL DL)</p><p>The W3C-endorsed OWL specification includes the definitionof three variants of OWL, with different levels ofexpressiveness.</p><p>OWL Lite was originally intended to support those usersprimarily needing a classification hierarchy and simpleconstraints. For example, while it supports cardinalityconstraints, it only permits cardinality values of 0 or 1. It washoped that it would be simpler to provide tool support for OWLLite than its more expressive relatives, allowing quickmigration path for systems utilizing thesauri and othertaxonomies. In practice, however, most of the expressivenessconstraints placed on OWL Lite amount to little more thansyntactic inconveniences: most of the constructs available in</p><p>OWL DL can be built using complex combinations of OWL Litefeatures. Development of OWL Lite tools has thus provenalmost as difficult as development of tools for OWL DL, andOWL Lite is not widely used.</p><p>OWL DL was designed to provide the maximumexpressiveness possible while retaining computationalcompleteness (all conclusions are guaranteed to becomputed), decidability (all computations will finish in finitetime), and the availability of practical reasoning algorithms.OWL DL includes all OWL language constructs, but they canbe used only under certain restrictions (for example, numberrestrictions may not be placed upon properties which aredeclared to be transitive). OWL DL is so named due to its</p><p>correspondence with description logic, a field of research thathas studied the logics that form the formal foundation of OWL.</p><p>OWL Fullis based on a different semantics from OWL Lite orOWL DL, and was designed to preserve some compatibilitywith RDF Schema. For example, in OWL Full a class can betreated simultaneously as a collection of individuals and as anindividual in its own right; this is not permitted in OWL DL.OWL Full allows an ontology to augment the meaning of thepre-defined (RDF or OWL) vocabulary. It is unlikely that anyreasoning software will be able to support complete reasoningfor OWL Full.</p><p>Each of these sublanguages is an extension of its simpler</p><p>predecessor, both in what can be legally expressed and inwhat can be validly concluded. Usually, Ontologies are definedto consist of abstract concepts and relationships (orproperties) only. In some rare cases, Ontologies are defined toalso include instances of concepts and relationships. Thefollowing three types of Ontologies are common in literatureand are classified on the basis of their generality:</p><p>Domain: they are domain-specific and are used captureknowledge in a particular domain, e.g., engineering, medicine,e-commerce, etc.</p><p>Generic: they capture general, domain-independent</p><p>knowledge (e.g., space and time). They are shared by largenumber of users across distinct domains. Examples areWordNet and CYC.</p><p>Application: they capture the knowledge necessary for aparticularapplication, e.g., ontology representing the structureof a particular web site .</p><p>In enterprises, Google and Yahoo!, the major web searchservices, are using ontology-based approaches to find andorganize the contents on the Web.</p></li><li><p>8/9/2019 Mapping Between Rdbms and Ontology a Review</p><p> 3/7</p><p>INTERNATIONAL JOURNAL OF SCIENTIFIC &amp; TECHNOLOGY RESEARCH VOLUME 3, ISSUE 11, NOVEMBER 2014 ISSN 2277-8616 </p><p>309IJSTR2014www.ijstr.org</p><p>Table 1`: Current Web and Semantic Web</p><p>Table 2: RDBMS Vs Ontology</p><p>4ONTOLOGY MAPPINGMichael Wick, Khashayar Rohanimanesh, Andrew McCallum,AnHai Doan [11] presented a fully supervised statistical modelfor ontology mapping based on conditional random fields. Thismodel accounts for uncertainty in both the data and the data'sstructure. Results on two domains and showed that oursupervised model is able to generalize across them has beenevaluated. Yuan An, Alex Borgida and John Mylopoulos [12]discussed about the different mapping methods from databaseto ontologies. Here, author focused on semi-automatic tool,called MAPONTO, that assists users to discover plausiblesemantic relationships between a database schema (relationalor XML) and an ontology, expressing them as logical</p><p>formulas/rules. Raji Ghawi and Nadine Cullot [13], focus on acomponent of the architecture which is a tool, called DB2OWL,that automatically generates ontologies from databaseschemas as well as mappings that relate the ontologies to theinformation sources. The mapping process starts by detectingparticular cases for conceptual elements in the database andaccordingly converts database components to thecorresponding ontology components. A prototype of DB2OWLtool is implemented to create OWL ontology from relationaldatabase. Table 3 depicts the comparative study of variousapproaches to convert database management system toontology. Mostafa E. Saleh [15], presented an approach forsemantic query in traditional relational database by</p><p>establishing ontological layer. In this paper, author has beendescribed following rules for converting database to ontology:</p><p>a) If the primary key of more than one relation is the samethen they should be merged in one ontological classand their attributes should be merged.</p><p>b) If the primary key of one relation is unique for thatrelation, and not contain the primary key in anotherrelation, then that relation will be considered as oneontological class.</p><p>c) If the foreign key in a relation Ri is a primary key inanother relation Rj, then there is an object property(named by its name in Ri) from Ri to Rj, and the domain</p><p>is Ri, and range is Rj.d) If the relation primary key consists of two other primarykeys, then that relation is a property between twoclasses (resources), the classes are the two relationsdenoted by the two primary keys.</p><p>Wei Hu and Yuzhong Qu [16], propose a new approach todiscovering simple mappings between a relational databaseschema and ontology. It exploits simple mappings based onvirtual documents, and eliminates incorrect mappings viavalidating mapping consistency. Man Li, Xiao-Yong Du, ShanWang [17], described the learning rules from relationadatabase to OWL ontology. In this paper, an ontology learningapproach has been proposed to construct OWL ontology</p></li><li><p>8/9/2019 Mapping Between Rdbms and Ontology a Review</p><p> 4/7</p><p>INTERNATIONAL JOURNAL OF SCIENTIFIC &amp; TECHNOLOGY RESEARCH VOLUME 3, ISSUE 11, NOVEMBER 2014 ISSN 2277-8616 </p><p>310IJSTR2014www.ijstr.org</p><p>automatically based on data in relational database. Therelated learning rules are discussed in detail. It can be seenthat the approach is practical and helpful to the automation ofontology building. Guntars Bumans [18], demonstrated on asimple yet completely elaborated example how mappinginformation stored in relational tables can be processed usingSQL to generate RDF triples for OWL class and propertyinstances. Noreddine Gherabi, Khaoula Addakiri [19],</p><p>prototype has been implemented, which migrates a RDB intoOWL structure, for demonstrate the practical applicability ofapproach by showing how the results of reasoning of thistechnique can help improve the Web systems. Authors havepresented a new approach for mapping relational databaseinto Web ontology. It captures semantic information containedin the structures of RDB, and eliminates incorrect mappings byvalidating mapping consistency. Secondly, we have proposeda new algorithm for constructing contextual mappings,respecting the rules of passage, and integrity constraints.Fuad Mire Hassan, Imran Ghani, Muhammad Faheem,Abdirahman Ali Hajji [20], in this authors has been reviewedand presented a number of articles for Human ResourceOntology in eRecruitment domain. The papers described the</p><p>human resource ontology used within ontology matchingapproach, which provides means for semantic matchingapproach to match job seekers and job advertisements in arecruitment domain. Marc Ehrig and Steffen Staab [21],considered QOM, Quick Ontology Mapping, as a way to tradeoff between effectiveness (i.e. quality) and efficiency of themapping generation algorithms and demonstrated that QOMhas lower run-time complexity than existing prominentapproaches. Jess Barrasa, scar Corcho, Asuncin Gmez-Prez [22], in this paper authors has been illustrated R2O, anExtensible and Semantically Based Database to-ontologyMapping Language. Authors presented R2O, an extensibleand declarative language to describe mappings betweenrelational DB schemas and ontologies implemented in RDF(S)</p><p>or OWL. R2O provides an extensible set of primitives with welldefined semantics. Michal Laclavk [23], presented theapproach for creating semantic metadata from relationaldatabase data. When building ontology based informationsystems, it is often needed to convert or replicate data fromexisting information systems such as databases to theontology based information systems, if the ontology basedsystems want to work with real data. RDB2Onto convertsselected data from a relational database to a RDF/OWLontology document based on a defined template. CarlosEduardo Pires, Damires Souza, Thiago Pachco, Ana CarolinaSalgado [24], has been presented a tool SemMatcher, formatching ontology-based peer schemas, combining differentmatching strategies (e.g....</p></li></ul>