A Predictive Model for the Temporal Dynamics of Information - - PowerPoint PPT Presentation

A Predictive Model for the Temporal Dynamics of Information Diffusion in Online Social Networks

Adrien Guille*, Hakim Hacid** *ERIC Lab, Université Lumière Lyon 2 **Alcatel-Lucent Bell Labs France

International Workshop on Mining Social Network Dynamics

Held in conjunction with the 21st International World Wide Web Conference - Lyon, France - 16th April, 2012

• Online social networks are very powerful tools for

the spread of information

• 2010 Arab spring (Howard et al., 2011)
• 2008 U.S. presidential elections (Hughes et al., 2009)
• etc.
• The case of Twitter
• Explicit network
• Following & mentioning ties

Context

2 / 12

• A. GUILLE and H. HACID - A Predictive Model for the Temporal Dynamics of Information Diffusion in Online Social Networks - MSND/WWW’12

• A model capable of predicting the temporal

dynamics

• Input: a social network, an information topic

Output: the time-serie of the volume of tweets generated by the diffusion

Purpose of our proposal

3 Time Volume of tweets

• A. GUILLE and H. HACID - A Predictive Model for the Temporal Dynamics of Information Diffusion in Online Social Networks - MSND/WWW’12

/ 12

State-of-the-art

4

basis network prediction nature

• bject

Bakshy et al., 2011

passive links topological URL

Yang & Counts, 2010

active links topological topic

Galuba et al., 2010

passive & active links topological URL

Yang & Leskovec, 2010

non-graphical temporal hashtag

Our approach

passive & active links temporal topic

• Predictive models for diffusion on Twitter
• A. GUILLE and H. HACID - A Predictive Model for the Temporal Dynamics of Information Diffusion in Online Social Networks - MSND/WWW’12

/ 12

• Raw data
• Yang & Leskovec: 4,76.108 tweets
• Kwak et al.: followers graph

(1,47.109 edges)

• Preprocessing
• Extraction of the active ties
• Identification of sub-networks
• Identification of cascades

Data

5

• A. GUILLE and H. HACID - A Predictive Model for the Temporal Dynamics of Information Diffusion in Online Social Networks - MSND/WWW’12

/ 12

• AsIC principle (Saito et al., 2010)
• A network: the graph of followers
• Diffusion probabilities & time-delays
• Asynchronous Independent Cascades

Basis of the proposed model

6

• A. GUILLE and H. HACID - A Predictive Model for the Temporal Dynamics of Information Diffusion in Online Social Networks - MSND/WWW’12

/ 12

• Assumption
• The dynamics of the spreading process at the

macroscopic level is explained by interactions at a microscopic level between pairs of users and the topology of their interconnections.

• Dimensions

= 13 attributes for each set (u1,u2,topic,time)

Our approach

7 user I H dTR hM mR hK A u1 0,78 0,12 0,7 0,65 1 0,125 u2 0,23 0,12 0,23 1 0,10 1 0,33

• A. GUILLE and H. HACID - A Predictive Model for the Temporal Dynamics of Information Diffusion in Online Social Networks - MSND/WWW’12

/ 12

social semantic temporal

...

• Supervised classification task

Binary class: diffusion/non-diffusion

• Bayesian Logistic Regression (BLR)
• Results
• Social dimension only

precision rate of 79%

• Social + temporal + semantic dimensions

gain of 7%

• Global precision: 85%

Estimating diffusion probabilities

8

• A. GUILLE and H. HACID - A Predictive Model for the Temporal Dynamics of Information Diffusion in Online Social Networks - MSND/WWW’12

/ 12

• Input
• a network of users U = {u1,u2,...,un}
• an information i = {k1,...,kp}
• a subset of initialy informed users
• extension of AsIC
• fu1,u2(i,t) based on BLR
• ru1,u2 = (1-I(u2)).10
• t simulates the course of each day

Building a prediction engine

9

• A. GUILLE and H. HACID - A Predictive Model for the Temporal Dynamics of Information Diffusion in Online Social Networks - MSND/WWW’12

/ 12

user I dTR u1 0,78 0,45 u2 0,23 0,89 u3 0,98 0,19 mR 0,21 0,09 0,85

... ... ...

• Visualization

Information about the release date of the new iPhone

Evaluation on a prediction task

10 17,5 35 52,5 70 1 3 5 7 9 predicted real 7,5 15 22,5 30 1 3 5 7 E1 E2 Time-serie:

Time (days) Volume of tweets Time (days)

• A. GUILLE and H. HACID - A Predictive Model for the Temporal Dynamics of Information Diffusion in Online Social Networks - MSND/WWW’12

/ 12

Quantitative results

11 E1 E2 E3 E4 ALL Gain over dynamics 25,19 % 39,23 % 29,21 % 3,22 % 24,21 % Gain over volume 42,89 % 47,70 % 34,49 % 40,07 % 41,29 % Overall gain 34,04 % 43,46 % 31,85 % 21,65 % 32,75 %

• Gain of our approach over the 1-time lag predictor

(Yang & Leskovec, 2010)

• Two aspects:
• Relative error on predicting volume
• Relative error on dynamics
• A. GUILLE and H. HACID - A Predictive Model for the Temporal Dynamics of Information Diffusion in Online Social Networks - MSND/WWW’12

/ 12

• Initial assumption confirmed by the experiments

i.e. predicting diffusion from local properties

• Observation of a particular pattern
• Improve volume estimation

Conclusion

12 20 40 60 80 1 3 5 7 9 adjusted prediction (c=1,6) predicted real Time-series:

Volume of tweets Time (days)

• A. GUILLE and H. HACID - A Predictive Model for the Temporal Dynamics of Information Diffusion in Online Social Networks - MSND/WWW’12

/ 12

• P. N. Howard, A. Duffy, D. Freelon, M. Hussain, W. Marai, and M. Mazaid. Opening closed regimes,

what was the role of social media during the arab spring?, 2011.

• A. Hughes and L. Palen. Twitter adoption and use in mass convergence and emergency events.

International Journal of Emergency Management, 6(3), 2009.

• E. Bakshy, J. M. Hofman, W. A. Mason, and D. J. Watts. Everyone’s an influencer: quantifying influence
• n twitter. WSDM’11, 2011.
• J. Yang and S. Counts. Predicting the speed, scale, and range of information diffusion in twitter.

ICWSM’10, 2010.

• J. Yang and J. Leskovec. Modeling information diffusion in implicit networks. ICDM’10, 2010.
• W. Galuba, K. Aberer, D. Chakraborty, Z. Despotovic, and W. Kellerer. Outtweeting the twitterers -

predicting information cascades in microblogs. In Proceedings of the 3rd conference on Online social networks, WOSN’10, 2010.

• K. Saito, M. Kimura, K. Ohara, and H. Motoda. Selecting information diffusion models over social

networks for behavioral analysis. PKDD’10, 2010.

• H. Kwak, C. Lee, H. Park, and S. Moon. What is Twitter, a social network or a news media?, WWW’10,

2010.

References

• A. GUILLE and H. HACID - A Predictive Model for the Temporal Dynamics of Information Diffusion in Online Social Networks - MSND/WWW’12

• Visualization

Information about the launch of Googlewave

Annexes

10 20 30 40 1 3 5 7 9 10 20 30 40 1 3 5 7 9 Time-serie: predicted real First network Second network

Time (days) Volume of tweets Time (days)

• A. GUILLE and H. HACID - A Predictive Model for the Temporal Dynamics of Information Diffusion in Online Social Networks - MSND/WWW’12