LSDIS > Projects > SemDis > COI

Semantic Discovery: Discovering Complex Relationships in Semantic Web

A NSF Medium ITR project

Detecting Conflict of Interest (COI) using Semantic Associations

Collaborative work UGA & UMBC

Description: The goal is to detect potential conflict of interest by means of analysis on semantic associations. Using a subset of DBLP and FOAF

Take One: We used a subset of DBLP and a subset of FOAF. Both networks were integrated with an algorithm for entity reconciliation

Live demo of Conflict of Interest Detection

Take Two: We demonstrated scalability by using all of DBLP data and a much larger FOAF dataset (1 order of magnitude larger). We improved the COI detection algorithm by using more robust collaboration strength measures and by considering more relationships (e.g., same-affiliation, co-editorship)

Data Sources:

Source Code:
The source code is in Java. We used the Java-bindings of BRAHMS to load all the files (about 1GB). We claim that scalability is possible by using an average laptop (and probably the first to use BRAHMS in OSX).
Earlier prototyping was done using main-memory implementation of SemDis API. The change to BRAHMS was quite easy because its Java-bindings implement such API. The source code is available (show/hide)

The main-memory implementation of SemDis API uses Jena's ARP (RDF Parser). Hence, some jar files are required and should be obtained from their respective distributions as indicated in jars-list (show/hide)

Evaluation Datasets:
Our evaluation datsets consists of sets of accepted papers in several conference tracks (of WWW2006) and their respective Program Committee members. We ran our COI detection over these and manually verified a sample of the results to adjust our method. There were relatively few relationships passing through the foaf part of the dataset and then back to DBLP entities. Hence, we took a sample of 200 foaf:Person entities that have at least one foaf:knows relationship to verify that the detection of COI worked properly with FOAF data. These datasets are available in this list (show/hide)

Publications:
Semantic Analytics on Social Networks: Experiences in Addressing the Problem of Conflict of Interest Detection (15th International World Wide Web Conference, Edinburgh, Scotland, May 23-26, 2006)

The contact person for details/problems/questions/etc on this page is Boanerges Aleman-Meza (baleman@uga.edu)


This material is based upon work supported by the National Science Foundation under Grant No. IIS-0325464 titled "SemDis: Discovering Complex Relationships in Semantic Web". Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.