Proximity and Semi-metric Networks for a Collaborative and Recommender Web Service

Luis M. Rocha*, Tiago Simas* Andreas Rechtsteiner**, Mariella DiGiacomo***, and Richard Luce***

*School of Informatics and Cognitive Science Program
Indiana University
1900 East Tenth Street, Bloomington IN 47408

**Center for Genomics and Bioinformatics
Indiana University
Bloomington IN 47408

***Research Library
Los Alamos National Laboratory
Los Alamos, NM 87545

Citation: Rocha, L.M., T. Simas, A. Rechtsteiner, M. DiGiacomo, R. Luce [2005]. "MyLibrary@LANL: Proximity and Semi-metric Networks for a Collaborative and Recommender Web Service". In: Proc. 2005 IEEE/WIC/ACM International Conference on Web Intelligence (WI'05), IIEEE Press, pp. 565-571.

The full paper is available in a preprint pdf version


We describe a network approach to building recommendation systems for a Web service. We employ two different types of weighted graphs in our analysis and development: Proximity graphs, a type of Fuzzy Graphs based on a co-occurrence probability, and semi-metric distance graphs, which do not observe the triangle inequality of Euclidean distances. Both types of graphs are used to develop intelligent recommendation and collaboration systems for the MyLibrary@LANL web service, a user-centered front-end to the Los Alamos National Laboratory's digital library collections and Web resources.

Keywords: Information Retrieval, Recommendation Systems, Recommender Algorithms, Digital Libraries, Collaboration, Collaborative Environments, Web, Networks, Network Science, Graph Theory, Fuzzy Graphs, Fuzzy Set Theory, Proximity, Probabilistic Network.

1. Introduction

The Web is used today as a means to integrate many electronic information resources. In particular, it enables the creation of personalized and collaborative digital library services. Indeed, the Web has changed the nature of scientific research by creating new expectations for libraries supporting research. Several digital library initiatives offer customized digital library environments, however, these services typically do not provide users with personalized and collaborative environments. MyLibrary at the Los Alamos National Laboratory (LANL) provides scientists with a personalized Web environment enhancing scientific collaboration independent of time and location. One of the unique characteristics of this capability is the ability to push recommendations to users and adapt the system further based on user interactions.

We have described some of the adaptive features of MyLibrary@LANL in other publications. In this paper we present a network analysis methodology to produce recommendation systems and enhanced collaboration in this particular web service. This methodology is applicable to other types of web services beyond digital libraries, as also discussed in this paper.

For the full paper please download the pdf version

For more information contact Luis Rocha at
Last Modified: September 26, 2005