Structural Feature Based Anomaly Detection for Packed Executable Identification - Computational Intelligence in Security for Information Systems - page 238

Produre: A Novel Proximity Discovery Mechanism in
Location Tagging System

Jianbin Hu, Yonggang Wang, Tao Yang, Zhi Guan, Nike Gui, and Zhong Chen

Key Laboratory of High Confidence Software Technologies, Ministry of Education, China
School of Electronics Engineering and Computer Science, Peking University, China
{hjbin,wangyg,yangtao,guanzhi,guink,chen}@infosec.pku.edu.cn

Abstract. Proximity discovery is a very interesting and useful technique which
helps a user to find out his proximities who have the same or similar location
with him during a certain period of time. However, current methods of discover-
ing proximities are difficult to adopt and vulnerable when wrong location tags are
provided. This paper proposes Produre, a novel proximity discovery mechanism
in location tagging system based on users' credibility degrees and online social
network. The introduction of online social network helps to maintain the rela-
tionship and location exposure policies between friends. The users' credibilities
help to diminish the bad influence of malicious users who always annotate wrong
location tags because the credibility scores change based on users' performance.
The experimental results by our prototype illustrate Produre can effectively dis-
cover proximities of a user in an efficient and accurate way with high quality and
fewer mistakes.

Keywords: Proximity discovery, location tagging system, credibility degree, on-
line social network.

1 Introduction

With the development of online social networks(OSN), such as Facebook [1] in U.S.
and Renren [2] in China, more and more applications have emerged to provide personal-
ized and useful services. Among all these services, proximity discovery has attracted so
many people and is being discussed almost everywhere. Here we define the term “prox-
imity discovery” as “finding out the persons who have the same or similar geographical
location with a certain user”. In other words, according to proximity discovery, we can
find out the geographical neighbors with respect to a certain user. In practice, we usu-
ally refer to the proximity of a user u as a user p the distance between whom and u is no
more than a threshold D t . In this paper, we would like to use “online social network”
as an information pool which can provide useful information (e.g., friend relationship,
location tags, maps...) to support proximity discovery. We define “user” as a real person
behind an individual account in an online social network.
This work was supported in part by the NSFC under grant No. 60773163, No. 60873238,
No. 60970135 and No. 61003230, the National Significant Science and Technology Projects
under grant No. 2009ZX01039-001-001, as well as the PKU PY Project under grant No. PKU-
PY2010-005.