acm-header
Sign In

Communications of the ACM

ACM TechNews

Stanford Computer Scientists Learn to Predict Which Photos Will Go Viral on Facebook


View as: Print Mobile App Share: Send by email Share on reddit Share on StumbleUpon Share on Hacker News Share on Tweeter Share on Facebook
Photos often shared on Facebook: kittens.

Researchers at Stanford University have developed a method that allows them to predict whether a photo will go viral on Facebook.

Credit: Anna Cobb

Stanford University researchers have developed a method for predicting which photos on Facebook will go viral. Their method involves studying cascades, the term used to describe photos or videos being shared multiple times.

According to recent data provided by Facebook, only one in 20 photos posted on the social network gets shared even once, and just one in 4,000 gets more than 500 shares. The researchers were able to predict when a photo cascade would double in shares with 80-percent accuracy.

The researchers began by analyzing 150,000 Facebook photos, each of which had been shared at least five times. The researchers initially found that, at any given point in a cascade, there was a 50-percent chance the number of shares would double. The researchers then looked for variables that might help them predict doubling events more accurately than a coin toss, including the rate and speed at which photos were shared, as well as the structure of sharing. The algorithm became more accurate the more times a photo was shared, as photos shared hundreds of times had an accuracy rate of 88 percent.

From Stanford Report (CA)
View Full Article

 

Abstracts Copyright © 2014 Information Inc., Bethesda, Maryland, USA


 

No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account