![Researchers Predict Twitter Trends With 95% Accuracy [STUDY]](https://helios-i.mashable.com/imagery/archives/016nTbJAwX9YZb11Tyx7plg/hero-image.fill.size_1248x702.v1647021602.jpg)
[brightcove video="1939423913001" /]
Researchers at MIT say they've created an algorithm for Twitter that predicts trending topics better than the site's existing equation.
Associate Professor Devavrat Shah and student Stanislav Nikolov say their new algorithm predicts trending topics with 95% accuracy an average of 90 minutes faster than Twitter -- sometimes, as early as five hours before.
The algorithm combs through a large sample of tweets -- some that trended well, and some that didn't -- and compares the data to new information to see if there are any patterns. If new tweets look like older tweets that have trended, then there's a chance a new trend is being formed. Simple.
The equation could be applied to anything that changes over time, the researchers say, like the stock market, movie ticket prices or the duration of a bus ride. For Twitter, the data could prove beneficial to advertisers looking to market a certain topic or trend.
Check out the video above to learn more. What other trends -- on Twitter or otherwise -- would you like to see better predicted? Tell us what you think.