![What is a Hype and Where Can I Get One? Mathias Lux [email_address] Department for Information Technology, Klagenfurt University, Austria](https://image.slidesharecdn.com/power-laws-popularity-and-interestingness-1201983209881159-3/85/Power-Laws-Popularity-And-Interestingness-2-320.jpg)
Power Laws Popularity And Interestingness
- 1. ITEC, Klagenfurt University, Austria
- 2. What is a Hype and Where Can I Get One? Mathias Lux [email_address] Department for Information Technology, Klagenfurt University, Austria
- 3. What is this about … Power Laws & Pareto Distributions Just a Theory? Conclusions ITEC, Klagenfurt University, Austria by betta_design http://www.flickr.com/photos/betta_design/2200198472/
- 4. The Long Tail Common for certain distributions Zipf‘s Law Power Law Pareto Distribution In Web 2 Context Chris Anderson … ITEC, Klagenfurt University, Austria maitland 82 - http://www.flickr.com/photos/maitland82/346065497/
- 5. Zipf‘s Law Few events occur often, many occur rarely P n ~ 1/n a ... Frequency of the n th ranked item, a close to 1. Prominent examples Ranking of words in documents Ranking of cities and their size Ranking of movies and sold cinema tickets … and many more ITEC, Klagenfurt University, Austria
- 6. Zipf‘s Law Plot of the word frequency in Wikipedia Most popular: the, of, and from http://en.wikipedia.org/wiki/Zipf's_law ITEC, Klagenfurt University, Austria
- 7. Pareto Distribution 80:20 Rule Economics Continous (Zipf is discrete) Practical issues Time Management, … ITEC, Klagenfurt University, Austria
- 8. Power Law Made famous by Albert Barabasi Scale free networks (web, power supply, …) In-degree of web sites, etc. Defines actually a class of distributions f(x)=a*x^b + e Pareto and Zipf are part of the group ITEC, Klagenfurt University, Austria
- 9. How to detect a power law? Simple empirical tests Draw points on a log-log plot Is it a „straight line“? ITEC, Klagenfurt University, Austria
- 10. How to detect a power law? Statistical Means E.g. KS-Test, Chi-Square Test Open research issue … See e.g. Clauset, A., Shalizi, C.R., Newman, M.E.J.: Power-law distributions in empirical data. arXiv:0706.1062v1 (2007) ITEC, Klagenfurt University, Austria
- 11. A note on plots … ITEC, Klagenfurt University, Austria Taken from phun.org, tnx to enzo nadrag
- 12. A note on statistical means … http://www.phun.org/newspics/funny_friday/2538.jpg tnx to Enzo Nadrag ITEC, Klagenfurt University, Austria
- 13. Zipf, Pareto & Power Law: Conclusions They emerge when people are involved They have interesting characteristics Mean has virtually no information Area under the curve (cp. amazon’s long tail strategy) Power laws emerge somehow … Multiple generative models (preferntial attachement, memory kernels, etc.) No one knows for sure ITEC, Klagenfurt University, Austria
- 14. Is this just theory? Basically: YES! But there are related practical questions Are you using Flickr? How many “interesting” photos did you publish? How many views do your photos have? Imagine you publish a video on YouTube What are the chances that your video is a big hit? How to “help out” the process of getting a big hit? Can one distinguish between hit or flop? ITEC, Klagenfurt University, Austria
- 15. Is this just theory? (2) More related practical questions Do you have a website? How to “flat out” resource popularity? How select popular resources (e.g. for caching, adaptation, preprocessing)? ITEC, Klagenfurt University, Austria
- 16. Big hits on YouTube ITEC, Klagenfurt University, Austria © 2007 by Aigner Thomas and Oraze Manuel
- 17. Getting popular … Starting with the first view (user) Some other users find the same resource They point other to it Blogging, Digging, word of mouth Multiplicator of information – cp. Metcalfe’s law Number of views (users) “explodes” ITEC, Klagenfurt University, Austria
- 18. Some graphs … ITEC, Klagenfurt University, Austria Data from del.icio.us Shows bookmarks / day relative user count
- 19. Observations There is an initial bend in the curve The mean user # at the bend is rather small Around 50 There are outliers Google Video was doomed to be a success ITEC, Klagenfurt University, Austria
- 20. Conclusions If there is a bend … Chances are better for a big hit. Time is still an issue Slow start, long vs. short hype, etc. Resources without this bend: Better Chances that they are shelf warmers Decision support for portfolio adaptation ITEC, Klagenfurt University, Austria
- 21. The Flickr way Flickr defined “Interestingness” Patented combining views, comments, age, etc. Interesting photos are presented Users see new photos Not all photos (2.000 new / minute, checked Feb. 1 2008, ~ 11.oo UTC) They have no “big hit” ITEC, Klagenfurt University, Austria Kudos given to Horst Gutmann and Marian Kogler
- 22. The YouTube way Smaller resource data base than Flickr Around 45 videos a day (65.000 a day) But a lot more views (data Feb. 1st, 08) 73.245.607 for „Evolution of Dance“ 20 most viewed have > 30M views Not obvious counter strategy Might not (yet) be necessary ITEC, Klagenfurt University, Austria
- 23. Digg Assumption: Diggs also follow a power law Quite reasonable … How to avoid the Digg- effect? Digg has a mirror … ITEC, Klagenfurt University, Austria
- 24. Thanks ... ... for your attention You are interested? Then talk to me … ITEC, Klagenfurt University, Austria by Gexydaf http://www.flickr.com/photos/gexydaf/2208215419/
- 25. Mathias Lux Affiliation Klagenfurt University, ITEC Contact mathias @ juggle.at http://www.semanticmetadata.net ITEC, Klagenfurt University, Austria