Predicting the Popularity of Online Content Export

@article{citeulike:4187564, abstract = {We present a method for accurately predicting the long time popularity of online content from early measurements of user's access. Using two content sharing portals, Youtube and Digg, we show that by modeling the accrual of views and votes on content offered by these services we can predict the long-term dynamics of individual submissions from initial data. In the case of Digg, measuring access to given stories during the first two hours allows us to forecast their popularity 30 days ahead with remarkable accuracy, while downloads of Youtube videos need to be followed for 10 days to attain the same performance. The differing time scales of the predictions are shown to be due to differences in how content is consumed on the two portals: Digg stories quickly become outdated, while Youtube videos are still found long after they are initially submitted to the portal. We show that predictions are more accurate for submissions for which attention decays quickly, whereas predictions for evergreen content will be prone to larger errors.}, author = {Szabo, Gabor and Huberman, Bernardo A.}, citeulike-article-id = {4187564}, citeulike-linkout-0 = {http://ssrn.com/abstract=1295610}, citeulike-linkout-1 = {http://papers.ssrn.com/sol3/Delivery.cfm/SSRN_ID1295610_code2040.pdf?abstractid=1295610&mirid=1}, comment = {"In this paper we presented a method and experimental verification on how the popularity of (user contributed) content can be predicted very soon after the submission has been made, by measuring the popularity at an early time. A strong linear correlation was found between the logarithmically transformed popularities at early and later times, with the residual noise on this transformed scale being normally distributed. Using the fact of linear correlation we presented three models for making predictions about future popular- ity, and compared their performance on Youtube videos and Digg story submissions. The multiplicative nature of the noise term allows us to show that the accuracy of the pre- dictions will exhibit a large dispersion around the average if a direct squared error measure is chosen, while if we take the relative errors the dispersion is considerably smaller. An important consequence is that absolute error measures should be avoided in favor of relative measures in community portals when the error of the prediction is estimated. " }, day = {05}, journal = {Social Science Research Network Working Paper Series}, keywords = {digg, huberman, popularity, prediction, socialweb, statistics, youtube}, month = {November}, posted-at = {2009-03-17 21:03:31}, priority = {2}, publisher = {SSRN}, title = {Predicting the Popularity of Online Content}, url = {http://ssrn.com/abstract=1295610}, year = {2008} }

TY - JOUR ID - citeulike:4187564 L3 - citeulike-article-id:4187564 N2 - We present a method for accurately predicting the long time popularity of online content from early measurements of user's access. Using two content sharing portals, Youtube and Digg, we show that by modeling the accrual of views and votes on content offered by these services we can predict the long-term dynamics of individual submissions from initial data. In the case of Digg, measuring access to given stories during the first two hours allows us to forecast their popularity 30 days ahead with remarkable accuracy, while downloads of Youtube videos need to be followed for 10 days to attain the same performance. The differing time scales of the predictions are shown to be due to differences in how content is consumed on the two portals: Digg stories quickly become outdated, while Youtube videos are still found long after they are initially submitted to the portal. We show that predictions are more accurate for submissions for which attention decays quickly, whereas predictions for evergreen content will be prone to larger errors. JF - Social Science Research Network Working Paper Series TI - Predicting the Popularity of Online Content PB - SSRN KW - digg KW - huberman KW - popularity KW - prediction KW - socialweb KW - statistics KW - youtube N1 - "In this paper we presented a method and experimental verification on how the popularity of (user contributed) content can be predicted very soon after the submission has been made, by measuring the popularity at an early time. A strong linear correlation was found between the logarithmically transformed popularities at early and later times, with the residual noise on this transformed scale being normally distributed. Using the fact of linear correlation we presented three models for making predictions about future popular- ity, and compared their performance on Youtube videos and Digg story submissions. The multiplicative nature of the noise term allows us to show that the accuracy of the pre- dictions will exhibit a large dispersion around the average if a direct squared error measure is chosen, while if we take the relative errors the dispersion is considerably smaller. An important consequence is that absolute error measures should be avoided in favor of relative measures in community portals when the error of the prediction is estimated. " AU - Szabo, Gabor AU - Huberman, Bernardo PY - 2008/November/05/ UR - http://ssrn.com/abstract=1295610 ER -