Top Banner
Active Opinion Maximization in Social Networks Xinyue Liu, Xiangnan Kong and Philip S. Yu Worcester Polytechnic Institute, University of Illinois at Chicago Product Like Like Like Like Like observed opinions on target product and other products Influence Diusion Expressed Opinion The Target Social Network budget=k 1 budget=k 2 Objective: Finding the optimal seed selection that maximizes the total opinion spread. Diusion Continues Seed Selection Next Round Seed Selection Update Update updated In round 1
2

Active Opinion Maximization in Social Networkscinv.ro/files/KDD18_blitz.pdf · 2019-01-21 · Active Opinion Maximization in Social Networks Xinyue Liu, Xiangnan Kong and Philip S.

Jul 28, 2020

Download

Documents

dariahiddleston
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Active Opinion Maximization in Social Networkscinv.ro/files/KDD18_blitz.pdf · 2019-01-21 · Active Opinion Maximization in Social Networks Xinyue Liu, Xiangnan Kong and Philip S.

Active Opinion Maximization in Social NetworksXinyue Liu, Xiangnan Kong and Philip S. Yu

Worcester Polytechnic Institute, University of Illinois at Chicago

Product

Like

Like

Like

Like

Like

observed opinions on target product and other products

Influence Diffusion

Expressed Opinion

The Target Social Network

budget=k1 budget=k2

Objective:Finding the optimal seed selection that maximizes the total opinion spread.

Diffusion Continues

Seed Selection

Next Round Seed

Selection

Update

Updateupdated

In round 1

Page 2: Active Opinion Maximization in Social Networkscinv.ro/files/KDD18_blitz.pdf · 2019-01-21 · Active Opinion Maximization in Social Networks Xinyue Liu, Xiangnan Kong and Philip S.

Contributions

• Problem Formulation (Active Learning in Social Influence)

• Methodology (Greedy Algorithm with Matrix Factorization)

• Link collaborative filtering with social influence diffusion

• Empirical study on real-world network data

Poster Session: Tuesday 7:00pm-9:30pm @ICC Capital Hall (Level 0)