An Efficient Alternative to Personalized Page Rank for Friend Recommendations

Document Type

Conference Proceeding

Publication Date

3-19-2018

Publication Title

Consumer Communications & Networking Conference (CCNC), 2018 15th IEEE Annual

Publisher

IEEE

First page number:

1

Last page number:

2

Abstract

In this paper, a new algorithm is proposed to calculate the value of friend score in a graph. The main goal is decreasing the running time and accuracy of calculation. Even though similar estimations have been done, the proposed method is faster because it increases the speed of computation with ignoring calculation of not involved nodes. This feature makes this method an acceptable candidate for large networks analyses like search engines and social media friend recommendation techniques. Based on the experiments done in this study, the FOF algorithm increases the speed of calculation by a factor of 28 on the two test datasets, namely, Amazon and Facebook.

Keywords

Friend recommendation; Online social networks; Page rank; Personalized page rank

Disciplines

Electrical and Computer Engineering

Language

English

UNLV article access

Find in your library

Share

COinS