Enhancing meta-portals using dynamic user context personalization techniques

doi:10.1016/j.jnca.2011.10.005

Journal of Network and Computer Applications

Volume 35, Issue 5, September 2012, Pages 1446-1453

https://doi.org/10.1016/j.jnca.2011.10.005 Get rights and content

Abstract

The Internet is flooded with information and the last decade its size has grown so many times that information search and presentation have become tedious tasks even for experienced users. Minor changes to existing resources can alter the situation and lead to major changes to the end user experience. In this manuscript we present the dynamic web personalization and document grouping infrastructure for meta-portals and the evaluation of our mechanism on a meta-portal. A meta-portal is an informational node where articles from different sources are collected and presented in a categorized and personalized manner. The web personalization mechanism is based on dynamic creation and update of user profiles according to the users preferences when browsing. In parallel a user's profile is affected by user grouping details, which are constructed by users with similar profiles. Assuming that required information, such as article tagging, keywords to categories matching and articles to categories relation is already part of the meta-portal we present a novel mechanism that can build and maintain a user profile which is formed without disturbing the user. Furthermore, we describe the real-time user-centred document grouping mechanism that is implemented to support the web personalization system and present the experimental evaluation of the whole system.

Introduction

The last decade can be inevitably referenced as the decade of dramatic changes to almost every aspect of our everyday life. The advances of technology are huge and the evolution of World Wide Web (Internet) can be recognized as enormous. This weird freedom that the Internet offers, attracts more and more people. More attractive is the fact that people are free to produce on-line content in an extremely easy way making thus the production of web content a trend. The Internet is a vast place of article production and it can be referenced without any doubt as a large newsletter. The problem that arises from the fact that the Internet becomes a place where the sources (media) are more than the consumers (readers) is that the customers are usually unable to locate useful information. By useful information we define the information that an user would like to be presented, without being disturbed by any other means of content.

Searching across the Internet through the wide variety of search engines could be a possible solution to the problem of locating information, but the outrageous number of results is uninviting. The search tools that exist within article's sources and the communication channels provided can be presented as a solution or even the ultimate solution; however, the user must “invent” these places before starting to use these services. Creating customized and personalized sections within web pages is another viable solution but some recent examples seem to become misleading for the plethora of different types of users that exist on the web. User personalization and user profiling seem to be the panacea of the current chaotic web status.

User personalization is usually conflicted with the term customization. The difference is vast as the customization refers to the structure and coloring of the web page, while personalization usually refers to the content itself. What we believe is that the user should be able to adapt not only the structure of a web page, but also the content that is presented. Talking about specific content, somebody can assert that the portals are taking measures towards this problem and the content is enriched with an indication about category and lately with tagging on articles. This is sufficient up to an extent but there is still much to be done in order to extend the portals so as to present user centred information. The solution could be found on user profiling and dynamic changes to the user profile according to his habits.

We present a novel mechanism for user profile construction and maintenance in meta-portals. Many worldwide known meta-portals are Yahoo¹ and Google news.² We enhance the operation of our meta-portal peRSSonal by providing dynamically changing user profiling features fully adapted on the user's needs and without need of any user input.

The rest of the paper is structured as follows. The next section presents the related work while the third section our system's architecture. In the fourth section the algorithmic analysis is presented and the following section includes sets of experiments. The manuscript is finalized with future work and concluding remarks on the implemented system.

Section snippets

Related work

Many efforts were presented in the latest years in order to provide a solution to the problem of user profiling within web sites or even across the Internet. There is a slight but enormous difference between user profiling (which leads to personalization) and customization of web sites. Customization is the capability that is provided to the user to alter the layout of a web site; which is the color, the font, the position of the elements, the order of the information and others. In the context

Architecture

The architecture of the system relies on distributed components which form the dynamic web user profiling system. We are putting the focus on the personalized profiling subsystem. We are also doing brief analysis of the other modules in order to cross-connect the features of our complete system, peRSSonal.³

The architectural schema consists of a series of subsystems, as depicted in Fig. 1. The collaboration between the distributed parts is

Algorithm analysis

The algorithm that fetches the article is very simple and is based on the fact that every web portal includes a series of RSS feeds that are offered to the end user. Opposed to having to visit every page of the many news portals that exist on the Internet, we fetch their RSS and more specifically the one that includes the daily “top stories”. From the XML structure of the feeds we can obtain the most important articles that are published to each news portals, together with information that have

Experimental evaluation

The experimental evaluation of the system consists of experiments conducted in order to present the creation and maintenance of the user profile and to provide information about the statistics of the document grouping procedure. We utilize peRSSonal meta-portal which we enhance with the dynamic user profiling mechanism and the document clustering subsystem, and we are executing our experiments on both real and virtual users that are registered officially in peRSSonal. In parallel we are

Conclusion and future work

In this paper we presented a mechanism that is able to complete a procedure of collecting news from news portals and blogs and present them personalized back to the end-users by applying furthermore document clustering algorithms. This mechanism is helpful for Internet users who are spending a considerable amount of time trying to locate news of their interest through major or minor news portals or even through RSS feeds (RSS readers). Despite the fact that the personalization micro-sites that

References (20)

C. Bouras et al.
PeRSSonal's core functionality evaluation: enhancing text labeling through personalized summaries
Data and Knowledge Engineering Journal
(2008)
B.J. Jansen et al.
Real life, real users and real needs: a study and analysis of user queries on the Web
Information Processing and Management
(2000)
S.S. Anand et al.
Intelligent techniques for Web personalization
R. Baraglia et al.
Dynamic personalization of web sites without user intervention
Communications of the ACM
(2007)
Bouras C, Kounenis G, Misedakis I, Poulopoulos V. A web clipping service's information extraction mechanism. In: 3rd...
Bouras C, Dimitriou C, Poulopoulos V, Tsogkas V. The importance of the difference in text types to keyword extraction:...
Casaola E. Profusion Personal Assistant: an agent for personalized information filtering on the Internet. Master's...
Garofalakis J, Giannakoudi Th, Vopi A. Personalized web search by constructing semantic clusters of user profiles. In:...
J.L. Herlocker et al.
Evaluating collaborative filtering recommender systems
ACM Transactions on Information Systems
(2004)
Z. Huang
A cube model for web access sessions and cluster analysis

There are more references available in the full text version of this article.

Cited by (5)

Intelligent patent recommendation system for innovative design collaboration
2013, Journal of Network and Computer Applications
Citation Excerpt :
This system adopts a layered distributed architecture, similar to DNS, which is different from current commercial search engines. Dong et al. (2011) present a novel semantic similarity model for describing the service ontology environment whereas Bouras and Poulopoulos (2012) propose a web personalization mechanism based on dynamic creation and automatic updates of user profiles to better match users preferences. This approach assumes that a user's profile is affected by other user's grouping details which are constructed with similar profiles.
Patents' search is increasingly critical for a company's technological advancement and sustainable marketing strategy. When most innovative designs are created collaboratively by a diverse team of researchers and technologists, patent knowledge management becomes time consuming with repeated efforts creating additional task conflicts. This research develops an intelligent recommendation methodology and system to enable timely and effective patent search prior, during, and after design collaboration to prevent potential infringement of existing intellectual property rights (IPR) and to secure new IPR for market advantage. The research develops an algorithm to dynamically search related patents in global patent databases. The system clusters users with similar patent search behaviors and, subsequently, infers new patent recommendations based on inter-cluster group member behaviors and characteristics. First, the methodology evaluates the filtered information obtained from collaborative patent searches. Second, the system clusters existing users and identifies users' neighbors based on the collaborative filtering algorithm. Using the clusters of users and their behaviors, the system recommends related patents. When collaborative design teams are planning R&D policies or searching patents and prior art claims to create new IP and prevent or settles IP legal disputes, the intelligent recommendation system identifies and recommends patents with greater efficiency and accuracy than previous systems and methods described in the literature.
Automated context dissemination for autonomic collaborative networks through semantic subscription filter generation
2013, Journal of Network and Computer Applications
Citation Excerpt :
In information systems, there is an increased attention towards context retrieval and dissemination research. While many solutions focus on effective presentation of the context (e.g., Bouras and Poulopoulos, 2012) another important challenge in context-related research is the dissemination of the context. Disseminating context in a scalable and effective way is key in designing a well performant information system.
The current manual management of services and applications in today's telecommunication networks is becoming increasingly complicated. In the Future Internet, management is assumed to be automated by introducing an autonomic layer of distributed management elements. These distributed management elements need to collaborate with each other to ensure end-to-end quality guarantees. In this paper, we focus on the context dissemination between such collaborative management elements. Context dissemination is the exchange of all relevant management data and knowledge between the elements. Collaborating elements typically generate large amounts of context and it is important to filter this continuous stream. We propose a context dissemination approach that automates the context exchange between elements. The approach enables the automated generation of semantic subscription filters. Subscription filters allow an element to define where, how, and when context needs to be requested from other entities. Moreover, the proposed approach allows making the subscription filter generation dependent on the context. We present algorithms that intelligently filter the knowledge that is stored in the ontology. The results show that the generation of subscription filters can be done in the order of tens of milliseconds.
Special issue on service delivery management in broadband networks
2012, Journal of Network and Computer Applications
Personalization of learning content and its structure, based on limitations of LMSs
2016, Studies in Computational Intelligence
Knowledge management and sharing in e-learning: Hierarchical system for managing learning resources
2014, KMIS 2014 - Proceedings of the International Conference on Knowledge Management and Information Sharing

View full text

Enhancing meta-portals using dynamic user context personalization techniques

Abstract

Introduction

Section snippets

Related work

Architecture

Algorithm analysis

Experimental evaluation

Conclusion and future work

Data and Knowledge Engineering Journal

Information Processing and Management

Intelligent techniques for Web personalization

Dynamic personalization of web sites without user intervention

Communications of the ACM

Evaluating collaborative filtering recommender systems

ACM Transactions on Information Systems

A cube model for web access sessions and cluster analysis