     |
| |
Twitter's pretty often enhancing its microblogging platform by adding plenty of new features, and the latest changes could very well be among the most useful ones. Via its Engineering Blog, Twitter has gone into detail about how it has improved search within its social network, noting that a "real-time computation" system has been built to help "identify search queries as soon as they're trending, send these queries to real humans to be judged, and then incorporate the human annotations into our back-end models." What this means, essentially, is you'll now be able to learn more about what the service's millions (and millions) of users are talking about, helping you discover what the most popular topics are all about -- you know, something like #notimpressed or #tigerblood. The company's blog post is rather thorough on how the search improvements work, so be sure to hit the source link below if you'd like to learn a little more about the recent tweaks.
Filed under: Internet
Source: Twitter
Even though there's an ongoing kerfuffle between social giants Instagram and Twitter, the microblogging service is putting that aside and focusing its own efforts on improving the experience for some of its (many) users. With that in mind, Twitter announced earlier today that it has expanded its trending topics feature, also known as Trends, to an additional 100 cities around the world, including big-name places like Frankfurt in Germany and Guadalajara down in Mexico. Of course, this means more people can now easily glance at some of the most talked about things around the Blue Bird social network -- still, don't expect Justin Bieber's name to pop up every single day on the list, as Twitter was pretty clear on how it feels about that.
Filed under: Internet
Via: The Next Web
Source: Twitter
Predicting the future of Twitter's trending topics is, as of right now, an impossibility. But two folks at the Massachusetts Institute of Technology may have cracked the code with an algorithm they're saying predicts -- with 95 percent accuracy -- the topics that will trend in the next hour and a half. The prediction has even been calculated as high as four to five hours ahead of time with the same level of accuracy. Not too bad!
Of course, beyond impressing friends with the predictions, the algorithm has direct implications for the likes of Twitter itself -- being able to sell ads against trending topics could benefit the social media company enormously in its ongoing quest to monetize. At any rate, it's distinctly less dangerous sounding than the last idea we heard involving Twitter and predictions. The algorithm will be presented next week at MIT's Interdisciplinary Workshop on Information and Decision in Social Networks, should you wish to dig into the nitty gritty of the math behind the madness.
Filed under: Misc, Science, Software, Alt
MIT prof and student discover algorithm for predicting trending Twitter topics originally appeared on Engadget on Fri, 02 Nov 2012 10:21:00 EDT. Please see our terms for use of feeds.
Permalink
SlashGear | 
MIT | Email this | Comments