Hany's JCDL Doctoral Consortium

46
Detecting, Modeling, & Predicting User Temporal Intention in Social Media Hany M. SalahEldeen Old Dominion University Advisor: Dr. Michael L. Nelson JCDL ‘12 Doctoral Consortium

description

Hany SalahEldeen's Doctoral Consortium Presentation

Transcript of Hany's JCDL Doctoral Consortium

Page 1: Hany's JCDL Doctoral Consortium

Detecting, Modeling, & Predicting User Temporal Intention

in Social Media

Hany M. SalahEldeen Old Dominion University

Advisor: Dr. Michael L. Nelson

JCDL ‘12 Doctoral Consortium

Page 2: Hany's JCDL Doctoral Consortium

Michael Jackson Dies

Snapshot on: June 25th 2009 http://web.archive.org/web/20090625232522/http://www.cnn.com/

Page 3: Hany's JCDL Doctoral Consortium

Jeff tweets about it…

Published on: June 25th 2009 https://twitter.com/mdnitehk/status/2333993907

Page 4: Hany's JCDL Doctoral Consortium

Jenny is off the grid

Jeff’s friend Jenny was on a vacation in Hawaii for a month…

Page 5: Hany's JCDL Doctoral Consortium

Jenny starts catching up a month later

Read on: July26th 2009

When she came back she checked Jeff’s tweets and was shocked!

https://twitter.com/mdnitehk/status/2333993907

Page 6: Hany's JCDL Doctoral Consortium

Jenny follows the link on July 26th

CNN page on: July 26th 2009 http://web.archive.org/web/20090726234411/http://www.cnn.com/

Page 7: Hany's JCDL Doctoral Consortium

Jenny is confused!

• Implication:

– Jenny thought Jeff is making a joke about her favorite singer and she got mad at him

• Problem:

– The tweet and the resource the tweet links to have become unsynchronized.

Page 8: Hany's JCDL Doctoral Consortium

The Egyptian Revolution

Page 11: Hany's JCDL Doctoral Consortium

Some tweets are still intact…

https://twitter.com/miss_amy_qb/status/32477898581483521

Page 12: Hany's JCDL Doctoral Consortium

…and some lost their meaning with the disappearance of the images

Missing ? https://twitter.com/aishes/status/32485352102952960

https://twitter.com/omar_chaaban/status/32203697597452289

Page 13: Hany's JCDL Doctoral Consortium

The tweet remains but the shared image disappeared…

http://yfrog.com/h5923xrvbqqvgzj

Page 14: Hany's JCDL Doctoral Consortium

Cairo….we have a problem

• Implication:

– The reader cannot understand what the author of the tweet meant because the image is not available.

• Problem:

– The post is available but the linked resource (image) is completely missing.

Page 15: Hany's JCDL Doctoral Consortium

The Anatomy of a Tweet

Page 16: Hany's JCDL Doctoral Consortium

The Anatomy of a Tweet Author’s username

Other user mention

Tweet Body

Hash Tag Shortened URL to resource

Publishing timestamp

Social Post

Shared Resource

Interaction options

Page 17: Hany's JCDL Doctoral Consortium

3 URIs = 3 Chances to fail

Page 18: Hany's JCDL Doctoral Consortium

Explanation in MJ’s example

Page 19: Hany's JCDL Doctoral Consortium

… t1

t4

t2

t3 t5 t7 t8 t9 tn

t6

Page 20: Hany's JCDL Doctoral Consortium

User’s Temporal Intention

Share time Implicit Explicit

Click time Implicit Explicit

Engineering problem Solved by providing

tools

The Focus of our research

Out of our scope Purview of Facebook, Twitter, Google, …etc

Instrumented shortener

Instrumented web client

Page 21: Hany's JCDL Doctoral Consortium

Sometimes you want a previous version

The Correct Temporal Intention

CNN.com at the closest time to the tweet: 25th June 2009 ~ 7pm

Page 22: Hany's JCDL Doctoral Consortium

Sometimes you want the current version

The Correct Temporal Intention

In this case the current state of the press releases page

Page 23: Hany's JCDL Doctoral Consortium

Research Question

Can we estimate the users’ intention at the time of posting

and reading to predict and maintain temporal consistency?

Page 24: Hany's JCDL Doctoral Consortium

Research Goals

• Detect the temporal intention of the:

1. Author upon sharing time

2. The reader upon dereferencing time

• Model this intention as a function of time, nature of the resource, and its context.

• Predict how resources change with time and the intention behind

sharing them to minimize inconsistency.

• Implement the prediction model to automatically preserve

vulnerable social content that is prone to change or loss.

• Create an environment implementing this framework that

provides a smooth temporal navigation of the social web.

Page 25: Hany's JCDL Doctoral Consortium

Related Work • User’s Web Search Intention

– A. Ashkan ECIR ’09

– C. Lee AINA ‘05

– A. Loser IRSW ‘08

– L. Azzopardi ECIR ‘09

– R. Baeza-Yates SPIR‘06

– N. Dai HT ’11

• Commercial Intention – Q. Guo SIGIR ’10

– A. Benczur AIRWeb ’07

• Sentiment Analysis – G. Mishne AAAI ‘06

– J. Bollen JCS ‘11

• Access to Archives – H. Van de Sompel OR‘09

• Persistence of shared resources – M. Nelson D-Lib ‘02

– R. Sanderson OR’11

– F. McCown JCDL ‘07

• URL Shortening – D. Antoniades WWW ’11

• Tweeting, Micro-blogging and Popularity – S. Wu WWW ’11

– A. Java SNA-KDD ’07

– H. Kwak WWW ’10

• Social Networks Growth and Evolution

– B. Meeder WWW ’11

Page 26: Hany's JCDL Doctoral Consortium

BEGIN

PhD Defense

Read Literature Collect Datasets Analyze Archives Coverage Analyze Shortened URIs Prototype Application Analyze Shared Resources Persistence and Coverage

Analyze Contextual Intention

Create Intention-based dataset Extract Intention Features Train a Parametric Model to predict intention Evaluate, test, cross-validate the model Create a mockup application Extend the model to induce preservation Finish Writing the Dissertation

Current State

Dissertation Plan

Page 27: Hany's JCDL Doctoral Consortium

BEGIN

PhD Defense

Read Literature Collect Datasets

Analyze Archives Coverage Analyze Shortened URIs Prototype Application Analyze Shared Resources Persistence and Coverage

Analyze Contextual Intention

Create Intention-based dataset Extract Intention Features Train a Parametric Model to predict intention Evaluate, test, cross-validate the model Create a mockup application Extend the model to induce preservation Finish Writing the Dissertation

Dissertation Plan

Page 28: Hany's JCDL Doctoral Consortium

Estimating Web Archiving Coverage • Goal: Estimate how much of the public web is present in the public archives

and how many copies are available? • Action:

– Getting 4 different datasets from 4 different sources: • Search Engines Indices • Bit.ly • DMOZ • Delicious.

• Results: *

• Publications: – How much of the web is archived? JCDL '11

* Table Courtesy of Ahmed AlSum JCDL 2011

Page 29: Hany's JCDL Doctoral Consortium

BEGIN

PhD Defense

Read Literature Collect Datasets Analyze Archives Coverage

Prototype Application Analyze Shared Resources Persistence and Coverage

Analyze Contextual Intention

Create Intention-based dataset Extract Intention Features Train a Parametric Model to predict intention Evaluate, test, cross-validate the model Create a mockup application Extend the model to induce preservation Finish Writing the Dissertation

Dissertation Plan

Analyze Shortened URIs

Page 30: Hany's JCDL Doctoral Consortium

Shortened URI analysis • Goal: Have a better understanding of URI shortening and resolving,

understand the effect of time on this process and the correlation between the page’s features and characteristics, and its resolution.

• Action:

– Fresh Bit.lys

– Get hourly clicklogs, rate of change, social networking spread, and other contextual information

– Longitudinal study

• Evaluation:

– Compare results with frequency of change analysis of Cho and Garcia-Molina.

– Compare results with Antoniades et al. WWW 2011.

Page 31: Hany's JCDL Doctoral Consortium

BEGIN

PhD Defense

Read Literature Collect Datasets Analyze Archives Coverage

Prototype Application

Analyze Shared Resources Persistence and Coverage

Analyze Contextual Intention

Create Intention-based dataset Extract Intention Features Train a Parametric Model to predict intention Evaluate, test, cross-validate the model Create a mockup application Extend the model to induce preservation Finish Writing the Dissertation

Dissertation Plan

Analyze Shortened URIs

Page 32: Hany's JCDL Doctoral Consortium

Estimating Loss of Shared Resources in Social Media

• Goal: Estimate how much of the public web is present in the public archives and how many copies are available?

• Action:

– Sampling from 6 public events

– Events spanning 3 years

– Existence in the current web

– Existence in the public archives

– Find relation with time

• Results:

– After 1st year ~11% will be lost

– After that we will continue on losing 0.02% daily

• Publications:

– A year after the Egyptian revolution, 10% of the social media documentation is gone. http://ws-dl.blogspot.com/2012/02/2012-02-11-losing-my-revolution-year.html

– Losing my revolution: How Many Resources Shared on Social Media Have Been Lost? TPDL '12

Page 33: Hany's JCDL Doctoral Consortium

BEGIN

PhD Defense

Read Literature Collect Datasets Analyze Archives Coverage

Prototype Application

Analyze Shared Resources Persistence and Coverage

User Intention Analysis

Create Intention-based dataset Extract Intention Features Train a Parametric Model to predict intention Evaluate, test, cross-validate the model Create a mockup application Extend the model to induce preservation Finish Writing the Dissertation

Dissertation Plan

Analyze Shortened URIs

Page 34: Hany's JCDL Doctoral Consortium

User Intention Analysis • Goal: Have a better understanding of User Intention and what factors affect

it. Also create a new testing and training set.

• Action:

– Get a sample set of tweets selected at random

– Extract the URIs

– Get closest Memento

– Download the snapshot & current version

– Use Amazon’s Mechanical Turk in choosing the best version

• Evaluation:

– Measure cross-rater agreement and confidence.

Page 35: Hany's JCDL Doctoral Consortium

Proposed Work

• Data Gathering

• Feature Extraction

• Modeling the intention engine

• Evaluation

• Application: Prediction and Preservation

Page 36: Hany's JCDL Doctoral Consortium

Possible Solution for Jenny

Page 37: Hany's JCDL Doctoral Consortium

Possible Solution for Jenny

The resource has changed since last time it was shared

Do you wish to see the version the author intended or the current version?

Current Version Intended Version

Page 38: Hany's JCDL Doctoral Consortium

Current Version

Archived Version

Proposed Framework

Feature Extraction

Classifier

Example Features: - Tweet Content - Click Logs - Other Tweets - Shared Resource - Timemaps

Page 39: Hany's JCDL Doctoral Consortium
Page 40: Hany's JCDL Doctoral Consortium

Extra Slides

Page 41: Hany's JCDL Doctoral Consortium

Archive Shortener Application

Page 42: Hany's JCDL Doctoral Consortium

Estimating Shared Resources Loss in Social Media

Page 43: Hany's JCDL Doctoral Consortium

Estimating Shared Resources Loss in Social Media

Page 44: Hany's JCDL Doctoral Consortium

My Publications

• S. G. Ainsworth, A. Alsum, H. SalahEldeen, M. C. Weigle, and M. L. Nelson. How much of the web is archived? In Proceedings of the 11th annual international ACM/IEEE joint conference on Digital libraries, JCDL '11, pages 133{136, 2011.

• H. SalahEldeen and M. L. Nelson. Losing my revolution: How much social media content has been lost? Accepted in TPDL 2012

• H. SalahEldeen and M. L. Nelson. Losing my revolution: A year after the Egyptian revolution, 10% of the social media documentation is gone. http://ws-dl.blogspot.com/2012/02/2012-02-11-losing-my-revolution-year.html.

Page 45: Hany's JCDL Doctoral Consortium

References • D. Antoniades, I. Polakis, G. Kontaxis, E. Athanasopoulos, S. Ioannidis, E. P. Markatos, and T. Karagiannis. we.b: the web of short

urls. In Proceedings of the 20th international conference on World wide web, WWW '11, pages 715 {724, New York, NY, USA, 2011. ACM.

• A. Ashkan, C. L. Clarke, E. Agichtein, and Q. Guo. Classifying and characterizing query intent. In Proceedings of the 31th European Conference on IR Research on Advances in Information Retrieval, ECIR '09, pages 578{586, Berlin, Heidelberg, 2009. Springer-Verlag.

• L. Azzopardi and M. de Rijke. Query intention acquisition: A case study on automatically inferring structured queries. In Proceedings DIR-2006, 2006.

• R. Baeza-Yates, L. Calderon-Benavides, and C. Gonzalez-Caro. The intention behind web queries. In F. Crestani, P. Ferragina, and M. Sanderson, editors, String Processing and Information Retrieval, volume 4209 of Lecture Notes in Computer Science, pages 98{109. Springer Berlin / Heidelberg, 2006. 10.1007/11880561 9.

• A. Benczur, I. Bro, K. Csalogany, and T. Sarlos. Web spam detection via commercial intent analysis. In Proceedings of the 3rd international workshop on Adversarial information retrieval on the web, AIRWeb '07, pages 89{92, New York, NY, USA, 2007. ACM.

• J. Bollen, H. Mao, and X.-J. Zeng. Twitter mood predicts the stock market. CoRR, abs/1010.3003, 2010.

• N. Dai, X. Qi, and B. D. Davison. Bridging link and query intent to enhance web search. In Proceedings of the 22nd ACM conference on Hypertext and hypermedia, HT '11, pages 17{26, New York, NY, USA, 2011. ACM.

• N. Dai, X. Qi, and B. D. Davison. Enhancing web search with entity intent. In Proceedings of the 20th international conference companion on World wide web, WWW '11, pages 29{30, New York, NY, USA, 2011. ACM.

• K. Durant and M. Smith. Predicting the political sentiment of web log posts using supervised machine learning techniques coupled with feature selection. In O. Nasraoui, M. Spiliopoulou, J. Srivastava, B. Mobasher, and B. Masand, editors, Advances in Web Mining and Web Usage Analysis, volume 4811 of Lecture Notes in Computer Science, pages 187{206. Springer Berlin / Heidelberg, 2007. 10.1007/978-3-540-77485-3 11.

Page 46: Hany's JCDL Doctoral Consortium

References • Q. Guo and E. Agichtein. Ready to buy or just browsing?: detecting web searcher goals from interaction data. In Proceedings of the 33rd

international ACM SIGIR conference on Research and development in information retrieval, SIGIR '10, pages 130{137, New York, NY, USA, 2010. ACM.

• A. Java, X. Song, T. Finin, and B. Tseng. Why we twitter: understanding microblogging usage and communities. In Proceedings of the 9th WebKDD and 1st SNA-KDD 2007 workshop on Web mining and social network analysis, WebKDD/SNA-KDD '07, pages 56{65, New York, NY, USA, 2007. ACM.

• H. Kwak, C. Lee, H. Park, and S. Moon. What is twitter, a social network or a news media? In Proceedings of the 19th international conference on World wide web, WWW '10, pages 591{600, New York, NY, USA, 2010. ACM.

• C.-H. L. Lee and A. Liu. Modeling the query intention with goals. In Proceedings of the 19th International Conference on Advanced Information Networking and Applications - Volume 2, AINA '05, pages 535{540, Washington, DC, USA, 2005. IEEE Computer Society.

• A. Loser, W. M. Barczynski, and F. Brauer. What's the intention behind your query? a few observations from a large developer community. In IRSW, 2008.

• F. McCown, N. Diawara, and M. L. Nelson. Factors aecting website reconstruction from the web infrastructure. In JCDL '07: Proceedings of the 7th ACM/IEEE-CS Joint Conference on Digital Libraries, pages 39{48, 2007.

• B. Meeder, B. Karrer, A. Sayedi, R. Ravi, C. Borgs, and J. Chayes. We know who you followed last summer: inferring social link creation times in twitter. In Proceedings of the 20th international conference on World wide web, WWW '11, pages 517{526, New York, NY, USA, 2011. ACM.

• G. Mishne. Predicting movie sales from blogger sentiment. In In AAAI 2006 Spring Symposium on Computational Approaches to Analysing Weblogs (AAAI-CAAW), 2006.

• M. L. Nelson and B. D. Allen. Object persistence and availability in digital libraries. D-Lib Magazine, 8(1), 2002.

• R. Sanderson, M. Phillips, and H. Van de Sompel. Analyzing the persistence of referenced web resources with memento. CoRR, abs/1105.3459, 2011.

• H. Van de Sompel, M. L. Nelson, R. Sanderson, L. Balakireva, S. Ainsworth, and H. Shankar. Memento: Time travel for the web. CoRR, abs/0911.1112, 2009.

• S. Wu, J. M. Hofman, W. A. Mason, and D. J. Watts. Who says what to whom on twitter. In Proceedings of the 20th international conference on World wide web, WWW '11, pages 705{714, New York, NY, USA, 2011. ACM.