A survey on ontology based web personalization


Published on

IJRET : International Journal of Research in Engineering and Technology is an international peer reviewed, online journal published by eSAT Publishing House for the enhancement of research in various disciplines of Engineering and Technology. The aim and scope of the journal is to provide an academic medium and an important reference for the advancement and dissemination of research results that support high-level learning, teaching and research in the fields of Engineering and Technology. We bring together Scientists, Academician, Field Engineers, Scholars and Students of related fields of Engineering and Technology.

Published in: Engineering
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

A survey on ontology based web personalization

  1. 1. IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308 __________________________________________________________________________________________ Volume: 02 Issue: 10 | Oct-2013, Available @ http://www.ijret.org 163 A SURVEY ON ONTOLOGY BASED WEB PERSONALIZATION Kiran Jammalamadaka1 , I V Srinivas2 1 Research Scholar, Computer Science, K L University, Andhra Pradesh, India, jvskkiran@gmail.com 2 Asst. Professor, Computer Science, K.J.SIMSR, Maharashtra, India, srinivasvj@gmail.com Abstract Over the last decade the data on World Wide Web has been growing in an exponential manner. According to Google the data is accelerating with a speed of billion pages per day [24]. Internet has around 2 million users accessing the World Wide Web for various information [25].These numbers certainly raise a severe concern over information over load challenges for the users. Many researchers have been working to overcome the challenge with web personalization, many researchers are looking at ontology based web personalization as an answer to the information overload, as each individual is unique. In this paper we present an overview of ontology based web personalization, Challenges and a survey of the work. This paper also points future work in web personalization. Index Terms: Web Personalization, Ontology, User modeling, web usage mining. ----------------------------------------------------------------------***------------------------------------------------------------------------ 1. WEB PERSONALIZATION Today, internet has become a part and parcel of our lives and one cannot imagine a world without internet, everyday millions of people use internet for various purposes mostly for information. And user is often not happy due the amount of information he has been provided with, as the user needs further filtering, which is very time consuming and expects the system to understand his/her thoughts. Understanding user is not as simple as it’s said, and web personalization is one step towards to the goal. Web personalization is the process of personalizing the content as per the user or set of users, taking advantage of the knowledge acquired through the analysis of the user’s navigational behavior [1]. Web personalization can be done in the following methods [17]. •Implicit: Implicit personalization will be performed by the system/web page based on the user behavior on the web •Explicit: User will be able to modify the system using the features provided by the system itself. •Hybrid: combination of the above. 1.1 Web Usage Mining In Personalization: Web Usage Mining is a subset of web mining which in turn a sub set of Data mining. Web mining has three categories Fig1. Web usage mining can be defined as automatic discovery of user profiles. The goal of web usage mining has been to support decision making process of website owners to understand the user in a better way. However, these techniques can be used for personalization functions. Web usage mining process consists of three major phases Fig2 Web usage mining is heavily dependent on the click stream data which gets generated automatically by the application servers in the form of logs.
  2. 2. IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308 __________________________________________________________________________________________ Volume: 02 Issue: 10 | Oct-2013, Available @ http://www.ijret.org 164 1.2 Need for Web Personalization Considering the amount of data and variety of users on the World Wide Web, key word based search results may not serve the purpose of providing the relevant information to the user, as each users’ intention is different and the same may not reflect in the key words they use. Because of the above reasons web personalization has attracted many researchers to look into and provide a mechanism to understand the user in a better way and provide most relevant information to the user. User may not have time to fill in the data (method Explicit) describing about his/her interests, likes, dislikes, background educational qualification etc. Many web mining researchers worked on the above challenge and provided a few techniques for automatic personalization, the best example till date was Amazon where user need not give his/her details the system will fetch the relevant information to the users. The below diagram explains overview of personalization techniques. [2]. Fig 3 Web personalization serves the purpose up to some extent by understanding the user however, relevant results cannot be provided to the user With-out understanding the context of the user. In the whole web personalization process, understanding the user and the context is the crux; towards it user modeling would be first step 1.3 User Modeling Assume a user is searching for a book title called “My experiments with truth”, then the user might be interested in autobiographies, in such case the personalized system will recommend many autobiographies to the user which could be apt for the user. However the user may not be an avid reader of autobiographies and the same user may get irritated by the very personalized system over the time. To avoid such scenarios the system needs to understand the user in a better way, certainly keywords without the context would not be much helpful. The field of user modeling has more than 35 years of research by now, User modeling has evolved from labs to commercial service models. ALFRED KOBSA [21]. in his paper reviewed the development of generic user modeling systems over the past twenty years. It described their purposes, services within user-adaptive systems, and the different design requirements for research prototypes and commercially deployed servers. The core idea of adaptation is based on the assumption that differences in some user characteristics should influence the individual utility of the service/information provided; hence if system’s behavior is tailored according to these characteristics, the system value will be increased. Some adaptive systems store individual information only for a single characteristic, others model users along multiple dimensions. [22]. A personalized system is said to be an adaptive system if the system is tuning itself to the requirement of the user, some personalized systems use domain knowledge which is Ontology. Ontology techniques will help user modeling in a better way. 2. ONOTOLOGY Ontology is a formal description and specification of knowledge. It provides a common understanding of topics to be communicated between users and systems [8]. As defined by Thomas R. Gruber as Ontology is "an explicit specification of a conceptualization”. A conceptualization consists of a set of entities (such as objects and concepts) that may be used to express knowledge and relationships [7]. Developing a Ontology includes
  3. 3. IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308 __________________________________________________________________________________________ Volume: 02 Issue: 10 | Oct-2013, Available @ http://www.ijret.org 165 Fig4 Ontologies have been proven an effective means for modeling digital collections and user context. Ontologies in the form of hierarchies of user interests have been proposed [11]. This ontology-based user modeling system integrates three ontologies: • User ontology: It includes different characteristics of users and their relationships. • Domain ontology: It captures the domain or application specific concepts and their relationships. • Log ontology: It represents the semantics of the user interaction with the system. [8]. The personalized ontology can describe different concept models for different users, although they may have the same topic. Ontology is based on two kinds of knowledge: 2.1 World Knowledge: World knowledge covering large number of topics so that the user’s individual information needs can be best match 2.2 Expert Knowledge: Expert knowledge is the kind of knowledge classified by the people who hold expertise in that domain. [9]. Ontologies are ever growing, constantly ontology repositories needs to be updated with the latest click stream data. Many ways in which ontology can be useful for user modeling have been discussed [23]. The approach is based on a semantic representation of the user activity, which takes into account both the structure of visited sites and the way the user navigates them. The most conventional implementation of ontology-based user modeling is the overlay user modeling that relies on a domain model represented as ontology. A simplified view on ontology is a network of concepts. [22]. Now days, Researchers identified the benefits of adapting of the Ontologies in user modeling. 3. SURVEYS Recently the need for adaptive systems has been recognized and the research has been geared up on user profiling and context. More sophisticated personalised web experiences have been provided with web mining techniques combined with Ontology. Robal and Kalja [12], discussed the possibilities of applying ontology’s in exploring the web sites’ structures and usage for producing various recommendations for the visitors. Content based system analyses item descriptions to identify items that are of particular interest to the user. Pazzani and Billsus [13] Discussed Content-based recommendation systems Kearney, Anand, & Shapcott [14], investigates how web visitor usage data may be combined with semantic domain knowledge to provide a deeper understanding of user behaviour. Seth and Zhang [15], focus on news associated content and suggest the design of a social network based recommender system for this rationale Staab & Studer [16] ,formally defined an ontology as a 4-tuple of a set of concepts ,a set of relations, a set of instances and a set of axioms. A context-aware system has to infer which context the user is in a given moment in time, and consequently adapt the system to that context [17]. The profiles are constructed using a variety of learning techniques including the vector space model Genetic algorithms, the probabilistic model or clustering. CASTELLS, Miriam FERNÁNDEZ, and David VALLET, Discussed about using Vector –space model for Ontology based information retrieval. [18]. C´elia da Costa Pereira1 and Andrea G. B. Tettamanzi, Discussed user profiling using genetic algorithms [19] Chen, Q. proposed a neural network approach to user modeling in the context of information retrieval [20]. 4. CHALLENGES 4.1 Accuracy of the User Profile: The biggest ever challenge is to build user model implicitly and accurately, as user may behave differently in each session some time it’s difficult to get a clear co relation of the
  4. 4. IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308 __________________________________________________________________________________________ Volume: 02 Issue: 10 | Oct-2013, Available @ http://www.ijret.org 166 navigation behavior. Reducing noise from the data collected is also a challenge. 4.2 Time Taking Process: User profile can be built over a period of time, until the system gets a minimum level of data, it may not provide the recommendations till such level. 4.3 Predicting & Updating the Users’ Model: Next challenge is predicting and updating the users’ behavior, a constant learning system needs to be in place. 4.4 Performance: Too much filtering can cause a performance issue which in turn leads the user to lose the interest. 4.5 Scalability: As the systems are dependent on Click stream data, log data etc., the systems should be capable enough to handle such large volumes of data. Though the processing can be done offline handling huge data is a challenge. 4.6 Privacy: User may not be comfortable in providing/allowing to capture his/her navigational data, it’s a serious privacy issue and government imposed many law restrictions on this, thus the systems may not get needed data from the user, Systems should be helping the users rather than spying. 4.7 User Model Interoperability: When the same user works with multiple adaptive systems, each system maintains its own recommendation system for the user, however it depends up on the quality of information each system has. The integration of these systems can definitely serve the user in a better way However; this task involves several important issues including privacy, security, trust user identity etc. [22]. 4.8 Identifying Web Robots: Web robots are also known as crawlers or spiders, which automatically traverse through the hyperlinks. Major challenge with web robots is that they hide their identity behind the user, and this may affect the user’s actual behavior, hence while building the user model the data generated might have the web robots’ influence. CONCLUSIONS The purpose of this survey has been to describe the state-of- the-art of personalized recommender systems using ontology techniques. In this paper, we have presented a comprehensive description about various web personalizations using ontology in the recent years. Though lot of research has been going on in this field, yet a system that effectively integrated various diverse requirements of the users has not yet been proposed. FUTURE WORK Future work in web personalization includes the in depth study on fusion of ontology and web mining techniques for effective web personalization. And how the user profiles evolves with time. This survey also identified a few areas to be explored like learning techniques including the vector space model, Genetic algorithms, and the probabilistic model or clustering in the field of web personalization. Integrating the systems like social networking and blogs and other popular websites is also a potential area to explore. Owing to the spread of mobile devices in the current era, web personalization needs to be explored on the mobile arena as well. REFERENCES [1]. [EV03] M. Eirinaki, M. Vazirgiannis, “Web Mining for Web Personalization”, ACM Transactions on Internet Technology, February 2003 [2]. Elizabeth Kulin, Rahul & Naureddien, “Web personalization Presentation” [Slides 5 & 18]. [3]. Chhavi Rana, “Trends in Web Mining for Personalization”, IJCST Vol. 3, Issue 1, Jan. - March 2012 , ISSN : 0976-8491 (Online) | ISSN : 2229-4333 (Print). [4]. Sergey Sosnovsky, “Ontological Technologies for User Modelling”, 2008 [5]. PLABAN KUMAR BHOWMICK, SUDESHNA SARKAR, ANUPAM BASU “ONTOLOGY BASED USER MODELING FOR PERSONALIZED INFORMATION ACCESS”, 2010 [6]. Mobasher, B.,“Data Mining for Web Personalization”. [7]. Gruber, T., “Toward Principles for the Design of Ontologies Used for Knowledge sharing”. 1993. [8]. K. Curran, C. Murphy, and S. Annesley, ” Web intelligence in information retrieval”. In Proc. of WI’ 03, pages 409 – 412, 2003. [9]. Xiaohui Tao, Yuefeng Li, Ning Zhong, Richi Nayak , “Ontology Mining for Personalized Web Information Gathering” [10].Natalya F. Noy and Deborah L. McGuinness, “Ontology Development 101: A Guide to Creating Your First Ontology” [11]. Joana Trajkova, Susan Gauch,” Improving Ontology- Based User Profiles “ [12]. Tarmo Robal, Ahto Kalja, ”Applying User Profile Ontology for Mining Web Site Adaptation Recommendations” [13]. J. Pazzani and Daniel Billsus, “Content-Based Recommendation Systems Michael “
  5. 5. IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308 __________________________________________________________________________________________ Volume: 02 Issue: 10 | Oct-2013, Available @ http://www.ijret.org 167 [14]. Sarabjot Singh Anand, Mary Shapcott, “Employing a domain ontology to gain insights into user behaviour Patricia Kearney†” [15]. Aaditeshwar Seth and Jie Zhang, “A Social Network Based Approach to Personalized Recommendation of Participatory Media Content” [16]. S. Staab and S. R., editors. “Handbook on Ontologies” [17]. E. Rich, “Users are individuals: individualizing user models”. Int. J. Man-Machine Studies (1983) 18, 199-214 [18]. Pablo CASTELLS, Miriam FERNÁNDEZ, and David VALLET, “An Adaptation of the Vector-Space Model for Ontology-Based Information Retrieval” [19]. C´elia da Costa Pereira1 and Andrea G. B. Tettamanzi, “An Ontology-Based Method for User Model Acquisition” [20]. Chen, Q, “A neural network approach for user modeling”. [21]. ALFRED KOBSA, “Generic User Modeling Systems” [22]. Sergey Sosnovsky, “Ontological Technologies for User Modeling” [23]. Hui Zhang, Yu Song, and Han-tao Song, “Construction of Ontology-Based User Model for Web Personalization” [24]. http://www.joop.in/Archive/the-world-wide-web-grows- a-billion-pages-per-day/ [25]. http://www.internetworldstats.com/stats.htm BIOGRAPHIES Kiran Jammalamadaka received his MCA from Acharya Nagarjuna University, India. He is currently associated with GE India. As a Software technical leader His Interests include Software engineering, web personalization, Agile->Scrum and Automated debugging I.V. Srinivas received his MCA from Acharya Nagarjuna University, India. He is currently associated with K. J. Somaiya Institute of Management Studies and Research, As a Asst. Professor. His interests include Distributed Computing, Software engineering and web personalization