Fuzzy Ontology Based Semantic Web Application Computer Science Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Email has become one of the fastest and most economical forms of communication. However Electronic Mails communication has been immensely adopted all around the world owing to the fact that it is one of the most feasible ,easily accessible and fastest modes of communication. E-Mails have provided huge amounts of impetuous towards the growth of the World Wide Web (WWW) network. Researchers recently have been actively involved in Semantic Web mail handling. The works carried out by most of the researchers are mainly oriented towards email with two main goals: Spam Mail filtering, and email topic classification, Spam & classification have more to do with keywords and not relations between emails.

This paper is aimed at identifying a Fuzzy Ontology based Semantic Web Application for Mail Servers (FOSM) to enable effective and efficient time management, workflow schedule management schemes, effective data understanding and incorporating ontology learning for the Arabic language. In the end, the researcher hopes not only to present a descriptive snapshot for the improvement of the current Fuzzy Ontology based Semantic Web Application for Mail Servers (FOSM) situation in a specific context, but also to give practicable recommendations that they can implement to increase its effectiveness in communications. Apart from this, the present study will be a valuable contribution to the growing body of work on the issue of effective e-mail communication. Thus, this study aims;

To realize a Fuzzy Ontology based Semantic Web Application for Mail Servers (FOSM) to enable effective and efficient time management and workflow schedule management schemes.

To realize these ontologies we propose to use Fuzzy Logic for effective data understanding.

To extend our research to incorporate ontology learning for the Arabic language too. Very few

Problem statement:

The problem we like to address and solve in this research is how to add semnantic layer to email server in order to make emails machine understandable .

Within this context, the research issues that we address are as follows:

How a Fuzzy Ontology based Semantic Web Application for Mail Servers (FOSM) will be effective and efficient in terms of:

time management?

workflow schedule management schemes?

What are the factors and steps to realize these ontologies to use Fuzzy Logic for effective data understanding?

How these research extends in order to incorporate ontology learning for the Arabic language?

Problem Structure:

This Section of the research proposal we present the work carried out in the following 3 areas

Semantic Web - Mail Servers

Fuzzy Based Ontology

Semantic Web support for Arabic Language

Each sub section is concluded based on the literature studied. Next 3 pages describe the three areas:

a. Semantic Web & Mail Server

The literary previous work on mail serve can be classified into three categories, Retrieval, Communication and Speech Act Theory.

McDowell et al (2004) investigated how the Semantic Web vision can be carried over to the realm of email. They also introduced a general semantic email notion, in which an email message is consists of a update or structured query together with explanatory text. They also suggested that semantic email opens the door to a wide range of email-mediated applications and automated with formally guaranteed properties. In particular, the researchers introduced a broad type of semantic email processes. They also showed that for the logical model it is really possible to infer automatically which email responses are w.r.t. acceptable set in polynomial time constraints , and for the model of decision-theoretic it is also possible for optimal message-handling computation of policy in polynomial time. In addition, the researcher show how to generate explanations automatically for a process's actions, and identifying cases where these type of explanations can be generated in terms of polynomial time. Finally, the researcher also discussed the publicly available semantic email implementation and outline research challenges in this sequences [13].

A study by Golbeck & Hendler (2004), assumed that in addition to traditional applications related to spam detection there are new filtering messages methods - including social network based filters and whitelist. These are being explored to further improvement in terms of mail classification and sorting [20].

In the article by Kassoff et al (2006), they explored how semantic technology can be delegated to addressing emails. They also introduced the semantic email addressing (SEA) notion that facilitates electronic mails to be delivered to a specified semantic recipient of recipient groupd, which may be change dynamically over time [21].

Yan and Dommel (2006) contributed specific factors on the combination of the Semantic Web poer, asynchronous collaboration and social networks, in a collaboration architecture. Their presentation also llustrated how the Semantic Web can be utilized to improved computer-supported collaboration using annotations in the Web and shareing application related to bookmarks [22].

Scerri et al (2007) tackle done of the biggest flaws of the communications through email - the insuffieciency of sharing expectations about the content and form of the interaction. According to the authors, this can be attributed to the insufficiency of explicit covering of the semantics context and exchanged email content that can be covered by applying speech act theory [23].

Youn and McLeod (2007) argued that increasing email users resulted in the elevated number of spam emails. As spammer tries to always find a way to evade existing e-mail filters and the necessity for a new emal filters arises to catch spam. According tho the authors, these ontologies facilitate machine-understandable data semantics significant to provide data with each other for more efficient filtering of spam emails [24].

Yang & Callan (2008) presented a new set of approach to identify different concepts showed in a email messages collection, and email organizing into a taxonomy or ontology for internet browsing. It came together with text mining techniques, natural language processing, information retrieval and machine learning to crerate an ontology concept Nominal N-gram mining Wordnet and surface text pattern were utilized [25].

Tafazzoli & Hadi (2008) introduces the term: malwareas malicious code written and designed to execute software system attacks. They reviewed the malact ontology and ontology of malware is also presented. It includes worms, spywares, backdoors, trojan horses, botnets, viruses, exploits and rootkits. In this paper they classified the malwares g based on their traits and attributes and presented a relationships in to 5 categories [26].

Kassim et al (2008) discussed the knowledge-based systems for email that focused on knowledge management incorporation approach for the enhancement of the traditional systems for e-mail. The author presented a knowledge based system of email called KS-Mail where individual is able to develop a sense of knowledge flow in their email exchangeing. They also introduced semantic e-mail processing through assigning email categories automatically and giving semantically links related to e-mail sent [27].

Riss el al. (2009) presented that applications for Email management were among the used mostly for collaboration tools in the enterprises. In illustrating the email and integrating task management idea and through technologies of semantic, they refered to a prototype system that they have created within the Nepomuk project framework. They also conducted interviews for a case study that undermine the expected advantages of three technologies combination [28].

Albakour & Blackwell (2009) utilized an approach to demonstrate how SemanticWeb and NLP techniques can be combined usefully to provide a practical applicationin a real-life with little customization manual. They also addressed a specic project management issue often located in huge collaboration of cross-enterprise projects that contains a various number of partners [29].

Chen (2010) presented a new face of an e-mail system from the speech act theory perspectives , and utilize the principle of six ethics of heart to formulate the social network kernel with provenance data to aid users in reducing the gap between their own personal management of information and current process of e-mail routine [30].

Zhang et al (2010) suggests that email's content is thoroughly reflected by semantic body being formed by a large semantic number elements. The researchers took the HowNet advantage in semantic element analysis and analyzed semantic email text bodies, then proposes the constructing semantic body method and the ways of calculation similarity between semantic bodies according to sentence similarity. Finally, they made a study of its application in filtering spam based on the proposed method. The experimental result showed that this method is more objective in email content judging, when it is being compared with the traditional method of email filtering in semantic unit. The proposed system reflected much better in discernment recall rate of email for spam emails whose meaning is unclearly expressed [14].

Scerri1 et al (2010) argued that email can be served as a virtual environment for working in which users or peinformation data belongs to workflows which is well-defined, these are largely unsupported and implicit by existing email clients. Semanta gives this support by Semantic Email enabling -enhanced email with machine-processable metadata about specific email types of Action Items. Semanta is presented with a bottleneck knowledge-acquisition, as users cannot be required to annotate each item action, and their automatically recognition proves to be difficult. Researchers focuses on computationally treatable application aspects of speech for the email action classification items. Finally, the system of rule-based will bootstrap a learning machine system that is in development currently, to form the initial sets of training which are then improved through the reviewing of the users [19].

Conclusion: It is very evident that a lot of work is being carried out to achieve and realize semantic mail servers. But it is evident that most work has been done towards Spam Filtering or email classification. Spam has more to do with keywords and not relations.

b. Fuzzy Based Ontology

In many application, Fuzzy Ontology domain is necessary to interpret vague data or information. This kind of ontology specifically aim for an efficient distributed and heterogeneous access to knowledge. Fuzzy ontology has been proposed to fix uncertainties and shortcomings and for further enhancement of information retrieval.

The work on Fuzzy ontology can be classified to three categories.

A study conducted by Stoilos et al (2006) suggests that the Semantic Web must handle applications that imprecise and face uncertain and information, including fuzzy, random, distorted and missing, knowledge. These researchers have proposed extending Description Logic and OWL to deal with such uncertainty. f-OWL, a fuzzyextension to OWL, can capture vague and imprecise knowledge. The accompanying Fuzzy Reasoning Engine facilitaes f-OWL reason and capture about such knowledge [12].

Calegari & Ciucci (2006) proposed that ontologies have proved to be very useful in terms of sharing concepts across different. The researchers also present that ontology-based applications information nowadays is often imprecise and vague. This is a common problem especially for semantics-based applications, such as knowledge management, e-commerce and web portals. In applications such as computer-aided reasoning, the predominant paradigm in managing vague knowledge is fuzzy set theory. They also present an classical computational ontologies enrichment with fuzzy logic to form fuzzy ontologies. They also proposal a developed application in the KAON ontology editor, that facilitates handling of ontology concepts in an high-level systems or environment [6].

Sanchez and Yamanoi (2006) presented in their studies entitled Fuzzy ontologies for the semantic webIt present a several connections between the Semantic Web, Fuzzy Logic and its components (Description Logics and Ontologies) and illustrated and introduced by an example ("Ontology of Art") a Fuzzy Ontology structure, Knowledge Base and Lexicon [39].

Tho & Hui (2006) also disccused ontology as an effective conceptualism used commonly for the Semantic Web. They mentioned that Fuzzy logic can be contrasted or associated to ontology in represent ing uncertainty information. Typically, the fuzzy ontology is being generated from a concept of predefined hierarchy. However, in constructing a concept hierarchy for a specific domain can be a tedious and difficult and task according to them. That is why, in tackling this issue, the researcher proposes the FOGA (Fuzzy Ontology Generation frAmework) for automatic fuzzy ontology generation on uncertainty information. The framework of FOGA comprises the components such as Concept Hierarchy Generation, Fuzzy Formal Concept Analysis, and Fuzzy Ontology Generation. They also discussed approximating reasoning for ontology's incremental enrichment with new upcoming data. Finally, they present a fuzzy-based technique for database attributes integration to the proposed ontology [40].

Gu, Lv & Gao (2007) discussed the current fuzzy ontologies and suggested that these are too domainspecific and cannot support fuzzy descriptions reasoning. They investigated the uncertainty information in the real world and discovered three existing wide general fuzzy relations. Based on their investigation and motivation to guide fuzzy ontology construction and support reasoning, they proposde a general reasoning-enabled fuzzy ontology including these generally fuzzy relations. In supporting such ontology in theory, thry introduce a fuzzy description logic ef-SHIN, and defined its semantics and syntax. They also presented a reference process for construction of fuzzy concept guidelines were based on relations in fuzzy ontology [6].

In the paper "Fuzzy Ontology and Information Access on the Web" GALLOVA (2007) assumed that the web is the largest repository of data available today. In her contribution for a solved Fuzzy set theory application of technique to the flexible systems definition and for accessing and locating information on the Web is presented. The researcher presented that the purpose the research is also a fact, that there are differemt ways to retrieve or access the mostly unknown information and big amount of available for users. Clustering methods are also proper for aiding on the document retrieval process and introduced a for solved system's fuzzy clustering methodology. The interface provided the users with necessary tools for navigating through documents and visualise in hierarchial form within the selected documents using the fuzzy clustering for HSC interface indexing [8].

Calegari & Ciucci (2006) presented a conceptual formalism supported by an ontology and suggested that it is not sufficient for handling or carrying vague information that is found commonly in many domain applications. They described how to introduce fuzziness in an ontology. And aimed to define a set of frameworks consisting of a fuzzy ontology based on Fuzzy---Owl and Fuzzy Description Logic [9].

Calegari & Sanchez (2007) showed how a Fuzzy Ontology based approach can semantic retrieval of documents. After defining a Fuzzy Knowledge Base formally, they discussed a special type of new relationships for non-taxonomic fuzzy, called (semantic) correlations. These type of correlations are updated after querying, or when a document has been inserted into a specific database. It is first assigned by experts. Information Retrieval algorithm is then introduced and allows to derive a path which is unique among the involved entities in the query in order to access maxima semantic association in the knowledge domain [10].

Lukasiewicz and Straccia (2008) argued that ontologies played a key role in the Semantic Web development through shared terms definition in web resources. The authors also showed that they are developed in web ontology languages, that are considerably on the basis of expressive logic description. They also gave approaches overview in this context for tha management of probabilistic uncertainty and vagueness in the Semantic Web's expressive logics description [31].

Zhai & Liang (2008) believed that information retrieval is a significant key role for Electronic Commerce. They argued that retrieval of Ontology-based semantic is a relevant subject for the current research. Their study applied a framework of fuzzy ontology to retrieval of information system in E-Commerce that include properties of concepts, concepts, and values of properties [32].

Trappey et al (2009) presented a hierarchical clustering novel approach for a self-organization knowledge document, specificall the forpatent analysis. They suggested that the approach for fuzzy logic control is utilized to match appropriate cluster(s) pf document based on their own ontological semantic webs [32].

De Maio et al (2009) disccessued that the success of Semantic Web is strongly related to the numerous diffusion of distributed ontologies enabling readable shared machine contents in the recent years. They also argued that ontologies vary in size, application domain, semantic, but often do not fore see the manipulation and representation of uncertain information. In this paper, they described t\\an approach for elicitation of automatic fuzzy ontology by the web resources collection analysis. The approach exploits a Formal Concept Analysis theory fuzzy extension and defines a system and methodological process in generating an OWL-based concepts, individual and properties representations of. This is considered as one of the simple case study in the Web domain that validates the flexibility and applicability of this approach [5].

Conclusion: Through our study we could conclude that Fuzzy Logic has been extensively used to derive meaning from the semantic web. Fuzzy Ontology is a very successful technique and highly results could be achieved hence we would incorporate Fuzzy Ontology techniques in our proposed framework.

Our fuzzy logic algorithm is scalable and would support Arabic too. The existing Fuzzy Ontology Learning techniques are predefined domain dependent, and are unable to extract relation between contents.

C. Semantic Web support for Arabic Language

There are studies conducted regarding Arabic language in Semantic Web. A proposed study was done to improve the information retrieval for Arabic Language on the Web in the legal domain by an Arabic search engine supporting the queries translation of Arabic language into French or English queries. The aim was to return documents written in Arabic, French or English [3].

Another research done on Arabic Word Net (AWN) based on the systems and methods developed for EuroWordNet (EWN) and since applied to dozens of languages around the world. This method approach maximizes compatibility across Word Nets and focuses on encoding of the most important and complicated concepts manually. The basic AWN criteria are relevance, connectivity, and generality, from English to Arabic and from Arabic to English [4].

Alrahabi & Helmy (2006) presented a semantic tool for annotation specifically used for Arabic texts with adaptedstrategy to the reported information's automatic location. The methodology utilized by the author was Contextual Exploration, which composed of utilizing purely knowledge about linguistics for the identification of representations for semantic-discursive text. Their work has been brought out in the a thesis framework at the University ParisIV-Sorbonne's LaLICC Laboratory [34].

Khalifa (2007) discuused that the Semantic Web advent in the late 90's, were lead to creation of many Web applications and benefited from the abilties presented by technologies relatied to Semantic Web. These technical abilites involve intelligent data reasoning, data interoperability and semantic search. However, most of these technologies related to Semantic Web are specifically dedicated for the Latin family scripts processing, thus, there is an apparent insufficiency of support related to Arabic script in these technologies. These isuues kept the Semantic Web research for the Arabic language actually untackled. The author reported a test on the support of Arabic experiment for the existing Semantic Web tools, and examined different opportunities and the challenges encountered for the application of Arabic Semantic Web and its implementation [35].

Belkredim1 & SebaiAn (2009) presented that there is a developemnt of complete and strong lexicon which is necessary for the processing of applications related to many Natural Language Processing (NLP) such as Question Answering Systems, Information Retrieval, Machine Translation and Natural Language Generation and. These lexicons according to the authors should be backed up by a reliable and strong theoretical frameworks and should be implemented utilizing different tools for the facilitation of efficient implementation and use of both language preciseness and in terms of resources. Recognition of the inflexional and derivative of the Arabic language nature, the authors presented a mathematics and ontologies based on formalism on the part of Arabic language' morphological knowledge. They also attempted to utilize the derivatives and their structural pattern for Arabic language and to linked the morphology of the words [36].

Hazman et al. (2009) said that ontologies serve an important role in many applications related to internet- and web-. Their work presented an accelerating system for the process of ontology building through a semi-automatically learning a hierarchal ontologies provided a set of seed concepts and a set of domain-specific documents in the web. The methodology were tested with web documents in the agriculture domain. The constructed ontology was based on the utilization of two approaches that complement with each other. They also presented a system that has been utilized for the building of a domain on agricultural ontology utilzing a set of documents for Arabic extensions and conductr and evaluation for a modified AGROVOC ontology version [37].

Bin Saleh (2009) discussed theachievement of the semantic web vision. He argued that web pages require to be changed into a form which can processable by a machine and enable the understanding of the computers. He also said that annotation of web page is considered as one of the techniques effective for knowledge and text artifacts representationThe author also presented a semantic annotation tool for Arabic called AraTation for annotating of Arabic News content semantically on the web constructed ustilizing the OWL ontologies and Java programming language for the production of for web pages' RDF metadata [38].

Abusalah et al (2009) reported an experiment for the evaluation of a Cross Language Information Retrieval (CLIR) system that utilizes a multilingual ontology in improving the query translation in the travel domain. The method of ontology-based approach is significantly outperformed the translation of Machine Readable Dictionary baseline utilizing the Mean Average Precision as usercentered in a metric experiment [16].

Qawaqneh & Eyas (2009) recently reported the numbers of Arabic web sites which were rapidly increasing in the internet. The existing search engines that retrieves information based on keywords is also increasing. This situation facilitates huge number of irrelevant data or information retrieved for users. By the the second generation appearance of the semantic World Wide Web the, many appropriate approaches in retrieving information depending on semantic can be generated. They also propsed that semantic web gives languages and data models such as Web Ontology Language (WOL) and Resource Description Framework (RDF) that facilitates building ontology within a particular domain. Moreover, this paper discussed the ontology system that can capture concepts for specific domain. Additionally, it can have reelationships of these concepts, and their properties. Therefore, they can facilitate machines to dealing with data domain semantically. The researcher proposed a newly designed approach in measuring the Arabic documents relevancy to the user query using concepts of ontology. Also on this paper, the researcher built concepts of Arabic ontology for the domain of electronic commerce in Arabic language [17].

Abouenour (2009) presented the semantic Query Expansion (QE) adoption that could be useful in the Question/Answering (Q/A) systems context . This is a challenging task for the Arabic language since it has many particularities such as absence of capital letters, short vowels, complex and morphology. The researchers presents a proposed semantic QE an evaluation based on Arabic WordNet (AWN). The researchers presented two different types of experiments which are conducted: the keyword-based evaluation which utilizes passage retrieval system ,a classical search engine and the evaluation of structure-based that utilizes the Java Information Retrieval System (JIRS) which carries into structure account of the question. The results showed that the excellent performances in terms of Mean Reciprocal Rank and accuracy that reached when using the proposed semantic QE with JIRS [18].

AI-Khalifa & AI-Yahya (2010) exhibited in their studies that the Holy Quran presents the classical Arabic's powerful rhetorical techniques; one of these methosds and techniques is the use of semantic opposition in the system. Semantic opposition is considered as a structural sub-field linguistics, studies between the relation of words based on a properties or set of features. In order to identify semantic opposition automatically terms in the Holy Quran, Semantic Web techniques such as ontologies can be utilised to represent semantic opposition. This paper discussed a work-in-progress building a framework project for identifying and recognizing semantic opposition in terms of using Natural Language processing associated with domain ontologies [15].

Conclusion: The support for Arabic Language for Semantic Web is very minimal. We would propose to overcome this feature in the research undertaken.