Social Search and Personalization Through Demographic Filtering

Introduction

With the growth of massive information on the Web, it has become increasingly difficult to search for useful information. As one of the most promising approaches to alleviate this overload, recommender systems have emerged in domains such as E-commerce, digital libraries, and knowledge management. In general, recommendation systems suggest items or products by analyzing what users with similar tastes have chosen in the past (Herlocker et al., 2002). There are two prevalent approaches to formulate recommendations: collaborative filtering recommendation and content-based recommendation. They depend on the type of items to be recommended and on the way that user models (Allen, 1990; Huang et al., 2004) are constructed. In collaborative filtering approach (Dalal, 2007; Herlocker et al., 2004; Hofmann, 2003), information is filtered for a larger group of users. The term collaborative filtering was coined by Goldberg et al. (1992). Collaborative-filtering algorithms aim to identify users that have relevant interests and preferences by calculating similarities and dissimilarities between user profiles. The idea behind this method is that it may be of benefit to one’s search for information to consult the behavior of other users who share the same or relevant interests. A content-based approach (Carmine, 2003; Kalles, 2003; Meng, 2002) provides recommendations by comparing an item’s content with the content that the user is interested in. In this approach, a model of user ratings is first developed. Then the filtering process is envisioned by computing the expected value of a user prediction given the user’s ratings on other items. Content-based algorithms are principally used when documents are to be recommended, such as web pages, publications, or news. The agent maintains information about user preferences either by initial input about user’s interests during the registration process or by rating documents. Recommendations are formed by taking into account the content of documents and by filtering in the ones that better match the user’s preferences and logged profile.

In this chapter, we propose an XML-based recommender system, called DemoFilter (Demographic Filtering). It is a type of collaborative information filtering system. DemoFilter uses ontology-driven social networks, where nodes represent social groups. A social group is an entity that defines a group based on demographic, ethnic, cultural, religious, age, or other characteristics. For example, people of ethnic group E_X; people who follow religion R_Y; and people who live in neighborhood N_Y can all be considered to form various social groups. In the DemoFilter framework, query results are filtered and ranked based on the preferences of the social groups to which the user belongs. If the user belongs to social group G_x, results will be filtered based on the preferences of G_x and the preferences of each ancestor social group of G_x in the social network. In social communities, it is commonly accepted that people who are known to share a specific background are likely to have additional connected interests (Herlocker et al., 2002). DemoFilter can be used for various practical applications, such as Internet or other businesses that market preference-driven products.