Research Article

Link Prediction Methods and Their Accuracy for Different Social Networks and Network Metrics

Table 1

Original dataset information.

Dataset name Time range Vertices Edges

Enron E-mail Communicationa 1998/11–2002/07 87,273 1,148,072
Facebook Wall Postsb 2008/01–2009/01 63,731 1,269,502
Flickr Friendshipc 2006/11–2007/05 2,302,925 33,140,018
PWr E-mail Communicationd 2008/11–2009/05 14,316 49,950
UC Irvine Messagese 2004/03–2004/10 1,899 59,835
YouTube Friendshipf 2006/12–2007/07 3,223,589 12,223,774

This table shows the original information about the datasets used in the experiments.
aThe Email network among employees of Enron. Nodes in the network are individual employees and edges are individual emails [32].
bThe wall posts from the Facebook New Orleans networks [33].
cThe social network of Flickr users and their friendship connections. It is collected by taking a snapshot of the network on November 2, 2006, and recording it daily until December 3, 2006, and then again daily between February 3, 2007, and May 18, 2007 [34, 35].
dThe Email Communication of Wrocław University of Technology [36].
eThe network contains messages sent between the users of an online community of students from the University of California, Irvine. A node represents a user. An edge represents sent message. Multiple edges denote multiple messages [37].
fThe social network of YouTube users and their friendship connections between December 10, 2006, and January 15, 2007, and again daily between February 8, 2007, and July 23, 2007 [38, 39].