Using ontology network analysis for research document recommendation

Abstract

Following tremendous advancement in information technology, the speed of information development has become increasingly fast-paced. Yet the overabundance of information has forced users to spend more time and resources in searching for information relevant to their needs. Today, recommendation systems already exist that provide services like filtering, customization, and others to assist users in searching for the right information. This study proposes to use ontology and the spreading activation model for research paper recommendation, hoping that it can elevate the performance of the recommendation system and also improve the shortcomings of today’s recommendation systems. This study utilizes ontology to construct user profiles and makes use of user profile ontology as the basis to reason about the interests of users. Furthermore, this study takes advantage of the spreading activation model to search for other influential users in the community network environment, making a study on their interests in order to provide recommendation on related information. Based on actual experiment results, the method of ontology network analysis that combines ontology and the spreading activation model is effective in knowing the research interests of users. Hence, using the mechanism proposed in this study can make up for the insufficiencies or shortcomings of other recommendation systems. Moreover, the precision rate can be up to 93% showing that our recommendation system has a positive effect on the effectiveness of the recommendation.

Introduction

Faced with the rapid advancement in information technology, organizations and enterprises have already properly equipped themselves with the right technology to create and store a large quantity of information. The information has, in turn, provided the important basis for the organization’s decision makers to set up marketing strategy, sales strategy, etc. Yet, in view of the overabundance of information, a phenomenon which is becoming serious and acute, many recommendation mechanisms have continuously tried to develop new methods from various theories that can assist their users to seek more relevant and more effective information.

In these last few years, ontology has gradually been widely developed and applied. Its definition lies in providing correct and explicit standards for shared concepts (Gruber, 1993). The merit of ontology lies in its ability to provide a clearly explained conceptual description of the relationships between entities in a specific domain. The Semantic Web concept proposed by Berners-Lee, Hendler, and Lassila (2001) is a result of the development of ontology to become a new generation Internet model.

In observing existing recommendation mechanisms or systems, admittedly the effects of information filtering of these systems have saved a lot of searching time for users. Yet this kind of recommendation method has failed to probe and to take advantage of the influential force and subsequent benefit of recommendations given by other users in the same network environment to the user. According to many studies, the most effective channel for delivery of information and knowledge is through informal organizations (Granovetter, 1983, Kraut et al., 1990, Wasserman and Galaskiewicz, 1994), for instance: through virtual communities or online social networks. On the other hand, Zhong, Liu, and Yao (2002) clearly pointed out that it is these kinds of communities or social network relationships that gave us the foundational information for our engagement in recommendation activities. As a result, if an enterprise is able to use this concept to deeply explore and analyze the hidden association between information, and through the combination of information sharing and the spreading mechanism, the enterprise can provide better and more suitable recommendations for customers. Not only this, by knowing the interests of specific groups, enterprises can create their marketing strategies, widening their customer bases and establish new business opportunities.

This study attempts to propose a framework to improve and upgrade the effectiveness of information recommendation. The primary objectives of this study are outlined below:

• Using ontology as the foundation for constructing user profiles.

  Using ontology on the construction of user profiles, the recommendation system is able to effectively create users’ concept hierarchies based on their interests, and automatically deduce and adjust itself to the changes in users’ likes, and does not need to go through frequent interactive activities with the users to learn about this.

• Using ontology network analysis to assist in recommendation.

  The strength of ontology network analysis lies in its ability to use its knowledge of specific domain to analyze the meaning of the information inside the network. Hence, it is suitable to be used to define concepts among groups. In addition, a weighted value can be assigned to it that will analyze the characteristics of groups and the network relationships in hope of exploring the relationships and values among the information.