Predicting Potential Responders in Twitter: A Query Routing Algorithm
0 likes352 views
A phenomenon not so recent is the substantial increase in popularity and use of online social networks. With that has emerged a new way to find information online: the social query, which consists of posting a question in a social network and wait for responses from close friends. Usually, a question is posted to be visible to everyone, but we believe that this is not the best way: there will be the possibility of receiving several responses (including wrong), keep receiving answers where there is no need, do not receive answers, etc. The query router problem consists of finding the most able individual in the personal social network of the questioner. This work presents an algorithm to Routing Questions in Twitter. The model was validated through its predict capacity and the results shows that its recommendations match in half cases only when combined with a technique to enrich the information present in the question.
![[DSC Europe 22] Machine learning algorithms as tools for student success pred...](https://cdn.slidesharecdn.com/ss_thumbnails/dijanaoreski-machinelearningalgorithmsastoolsforstudentsuccessprediction-221130080709-c0f224ba-thumbnail.jpg?width=560&fit=bounds)
More Related Content
Similar to Predicting Potential Responders in Twitter: A Query Routing Algorithm (20)
More from Jonathas Magalhães (8)
Recently uploaded (20)
Predicting Potential Responders in Twitter: A Query Routing Algorithm
- 1. Predicting Potential Responders in Twitter: A Query Routing Algorithm Cleyton Caetano de Souza Jonathas José Magalhães Evandro de Barros Costa Joseana Macêdo Fechine
- 2. Introduction • Online Social Networks (OSN) – Have became very popular – New way of using their virtual environments • Social Query – A new way to find information online – Publish a question to all your contacts [email protected] 2
- 3. What Kind of Questions? • Questions that are not well answered by Conventional Search Engines – Personal Questions – High Contextualized Questions – Recommendation Request – Opinion Request • Share your question with all your contacts • Wait for Answers (?) [email protected] 3
- 4. Problem • We believe that a public question is not the best strategy – Multiple Answers – Contradictory Answers – Wrong Answers – None Answers – Timeline Effect [email protected] 4
- 5. Solution • Direct the question to just one person – Ensures that the message will be viewed – But still, There are no guarantees about the quality of response • To whom should I direct questions? – The right one • Who is the right one? [email protected] 5
- 6. Features of the Right Person • Knowledge (𝐾) – He-She knows about the subject of the question • Trust (𝑇) – I trust that his-her answer will be truly • Availability (𝐴) – He-She will answer quickly [email protected] 6
- 7. Related work: About Ask Question in OSN • (Morris, Teevan and Panovich 2010a) – 93.5% of users received answers to their question after post them and these responses – in 90.1% of cases, were provided within one day [email protected] 7
- 8. Related work: About Ask Question in OSN • (Paul, Hong and Chi 2011) – 18.7% of questions posted on Twitter receive answers – 95% are answered within the range of 10 hours – the fact of receive or do not is intrinsically connected to the amount of followers of the questioner [email protected] 8
- 9. Related Work: About Direct Questions • Aardvark (Horowitz and Kamvar 2010) • iLink (Davitz et al 2007) • Q-Sabe (Andrade et al 2003) • AskWho (Liu 2010) [email protected] 9
- 10. Propose • A Routing Algorithm to route questions in OSN • Our Differential – A pre-existent social network as context – A flexible algorithm – A multi-criteria decision making problem • How evaluate a Routing Algorithm? [email protected] 10
- 11. Hypotheses • 𝐻0,1 : The proposed Routing Algorithm cannot predict the events of the real world at least 50% of trials; • 𝐻 𝑎,1 : The proposed Routing Algorithm can predict the events of the real world at least 50% of trials; [email protected] 11
- 12. Hypotheses • 𝐻0,2 : The proposed Routing Algorithm combined with the synonymy expansion in question cannot predict the events of the real world at least 50% of trials; • 𝐻 𝑎,2 : The proposed Routing Algorithm combined with the synonymy expansion in question can predict the events of the real world at least 50% of trials; [email protected] 12
- 13. Hypotheses • 𝐻0,3 : The combination of the Routing Algorithm with the synonymy expansion do not produces a recall rate higher than the same technique without expansion; • 𝐻 𝑎,3 : The combination of the Routing Algorithm with the synonymy expansion produces a recall rate higher than the same technique without expansion; [email protected] 13
- 14. The Model • Presented in Details in (Souza, Magalhães and Costa 2011) • The twitter is defined by the tuple 𝑇 = {𝑈, 𝑅} • Where 𝑈 = {𝑢1 , … , 𝑢 𝑈 } is a set of users • And 𝑅 is the set of all relationships 𝑟𝑖,𝑗 between two users 𝑖 and 𝑗. – The existence of 𝑟𝑖,𝑗 means that i follows j, this way 𝑟𝑖,𝑗 ≠ 𝑟𝑗,𝑖 [email protected] 14
- 15. The Model • Each useru has the attributes – 𝐹𝑜𝑙𝑙𝑜𝑤𝑒𝑟𝑠 𝑢 that contains all users which follows 𝑢 – 𝐹𝑜𝑙𝑙𝑜𝑤𝑖𝑛𝑔 𝑢 that contains all users which are followed by 𝑢 – 𝑀 𝑢 = 𝑚1 , … , 𝑚 𝑀 a ordered list that contains all messages posted for 𝑢 • Each message 𝑚 has the attributes – 𝑑 𝑚 - the post date – 𝑠 𝑚 - the string posted [email protected] 15
- 16. The Problem Given a query 𝑞 posted by 𝑢, 𝑓 ∈ 𝐹𝑜𝑙𝑙𝑜𝑤𝑒𝑟𝑠 𝑢 and 𝑝 𝑓,𝑞 a function that tell us the chances of 𝑓 provides a good answer – Find: 𝑓 – To: 𝑀𝑎𝑥 𝑝 𝑓,𝑞 – Over: 𝐹𝑜𝑙𝑙𝑜𝑤𝑒𝑟𝑠 𝑢 [email protected] 16
- 17. The Problem • We believe that 𝑝 𝑓,𝑞 has a correlation with three things – 𝑘 𝑓,𝑞 – the knowledge that 𝑓 in relation with 𝑞 – 𝑡 𝑢,𝑓 – the trust of 𝑢 has in 𝑓 – 𝑎 𝑓 – the level of activity of 𝑓 • That way will actually want to find the best combination of: 𝑘 𝑓,𝑞 , 𝑡 𝑢,𝑓 and 𝑎 𝑓 [email protected] 17
- 19. Evaluation • An experiment whose objective was to ascertain its ability to reflect, trough recommendations, what happened in real world • Nine volunteers posted on Twitter twenty nine questions which were answered fourth four users • The study involved the processing of a graph composed for 1201 users, 131.962 messages and 2.047.305 connections. [email protected] 19
- 20. Our Results Amount of True Positive without Expansion Amount of True Positive with Expansion 30 28 Amount of True Positive 25 24 22 22 24 20 21 18 18 15 11 10 10 5 4 0 1 5 10 15 20 25 Size of Recommendation List [email protected] 20
- 21. Our Results • The analysis over the recall rate indicated that hypotheses 𝐻0,1 and 𝐻 𝑎,2 was accepted. • Furthermore, the recall rate of both techniques were compared and the obtained conclusion is that the technique with synonymy expansion present results statically better than the simple technique (without expansion), confirming the hypotheses 𝐻 𝑎,3 . [email protected] 21
- 22. Conclusions • During the study, it was noted that the proposed task was naturally difficult • But, The fact that the recommendation match with what happens in the real world consists of a predictive validity of the conceptual model, but little refers to the quality of the recommendation. • These were preliminary results [email protected] 22
- 23. Future Work • A qualitative evaluation of the recommendations by the own questioner • A study on which factor is most important on the recommendation of experts: knowledge (𝑘 𝑓 𝑢,𝑞 ), trust (𝑡 𝑢,𝑓 𝑢 ) or activity (𝑎 𝑓 𝑢 ); and if its importance depends on the type/topic [email protected] 23
- 24. Future Work • If the direction of questions to a user (or a small number of users) is more effective than post the question to all followers. • Improve the results obtained by routing algorithm – Semantic Web Techniques – Bayes Theorem [email protected] 24
- 25. References • Andrade, J. C., Nardi, J. C., Pessoa, J. M. & Menezes, C. S. de. 2003. Qsabe-um ambiente inteligente para endereçamento de perguntas em uma comunidade virtual de esclarecimento. LA-WEB. • Davitz, J., Yu, J., Basu, S., Gutelius, D. & Harris, A. 2007. iLink: search and routing in social networks. 13th ACM SIGKDD International Conference on Knowledge discovery and data mining. • Horowitz, D. & Kamvar, S. D. 2010. The anatomy of a large-scale social search engine. 19th International Conference on World Wide Web. • Liu, C. (2010). AskWho: Finding Potential Answerers for Status Message Questions in Social Networks. agora.cs.illinois.edu. • Morris, M. R., Teevan, J. & Panovich, K. 2010. What do people ask their social networks, and why?: a survey study of status message Q&A behavior. 28th International Conference on Human factors in Computing Systems. • Paul, S. A., Hong, L., & Chi, E. H. (2011). Is Twitter a Good Place for Asking Questions? A Characterization Study. Fifth International AAAI Conference on Weblogs and Social Media. • Souza, C. C. D., Magalhães, J. J. & Costa, E. B. 2011. A Formal Model to the Routing Questions Problem In The Context Of Twitter. IADIS International Conference of WWW/Internet. [email protected] 25
- 26. Predicting Potential Responders in Twitter: A Query Routing Algorithm THANK YOU! Any Question?