Real-time search outfit OneRiotannounced today some updates to their search algorithm, which parses data in real-time social streams to index and rank links.
Although results based on freshness alone are available through the search engine’s real-time firehose setting, the results returned through the Pulse Rank setting are weighted based on several factors that riff off similar considerations for the static web and Google’s PageRank system.
Although Pulse setting searches have been around since OneRiot’s launch, the company announced today in their blog, “We made a couple of changes to our algorithm today to improve those results even further.”
Pulse results will now be a product of the following considerations: Timeliness, domain authority, individual reputation, and acceleration. The new algo attempts to weed out spam links (or links from known spammers) and give preference instead to “more thoughtful sharers whose links tend to get retweeted and dugg.”
According to the blog, the algo tweaks will also improve detection of whether a page is trending or whether it’s simply a large, popular destination and always has been.
“We’re also getting a lot of leverage from our Artificial Intelligence systems that constantly ‘learn’ how to improve the way we rank results,” the post continues.
