#algorithm "Google was one of the first major search engines to use a sophisticated algorithm to determine the relevancy of web pages in search results. Their ‘PageRank’ algorithm – together with a ton of other signals – became synonymous with Google and rendered many of the incumbents useless almost overnight. Goodbye Excite and Lycos. *** "Facebook doesn’t provide any information about how EdgeRank works (unlike Google which does) but the three commonly accepted factors at its core on are: Affinity – how close the relationship is between the user and the content/brand/source Weight – what type of action is being taken on the content, how important Facebook considers content type (video, photo, link etc) to be Decay – how recent the content is These legacy metrics now influence a myriad of complex calculations that produce a very personalised news feed. Last year, Lars Backstrom, Engineering Manager for News Feed Ranking at Facebook estimated that there are as many as “100,000 individual weights in the model that produces News Feed.” He says that “The easiest analogy is to search engines and how they rank web pages. It’s like comparing the Google of today with Alta Vista. Both Google and Bing have a lot of new signals, like personalization, that they use. It’s more sophisticated than the early days of search, when the words on a page were the most important thing.”"
Youtube personalization algorithm is "easily a million lines of code." "5 years ago (2009), ranking the search results by view counts may have been a great thing to do. Since then we've found other things that are even better signals." "If we just rank by view counts, you'd never see any new videos." How long a video has been watched on average is probably a better indicator of quality than how many times it has been viewed.
How do you create a good social algorithm: Some analysis, some intuition, and a lot of experimentation.
Hacker News ranking algorithm: Score = (P-1) / (T+2)^G where, P = points of an item (and -1 is to negate submitters vote) T = time since submission (in hours) G = Gravity, defaults to 1.8 in news.arc Effects of gravity (G) and time (T) Gravity and time have a significant impact on the score of an item. Generally these things hold true: the score decreases as T increases, meaning that older items will get lower and lower scores the score decreases much faster for older items if gravity is increased.
Reddit ranking algorithm: Submission time is a very important parameter, generally newer stories will rank higher than older. The first 10 upvotes count as high as the next 100. E.g. a story that has 10 upvotes and a story that has 50 upvotes will have a similar ranking. Controversial stories that get similar amounts of upvotes and downvotes will get a low ranking compared to stories that mainly get upvotes.
"... much of the time, there’s no way to tell why information is filtered the way it is online. Why is one person’s status update on Facebook prioritized in your News Feed over another’s? Why does Google return a different order of search results for you than for the person sitting next to you, googling the same thing? These are the mysteries of the algorithms that rule the web. And the weird thing is, they aren’t just inscrutable to the people clicking and scrolling around the Internet. Even the engineers who develop algorithms can’t tell you exactly how they work."
"... the majority illusion can occur in all of them. “The effect is largest in the political blogs network, where as many as 60%–70% of nodes will have a majority active neighbours, even when only 20% of the nodes are active,” they say. In other words, the majority illusion can be used to trick the population into believing something that is not true. That’s interesting work that immediately explains a number of interesting phenomena. For a start, it shows how some content can spread globally while other similar content does not—the key is to start with a small number of well-connected early adopters fooling the rest of the network into thinking it is common. That might seem harmless when it comes to memes on Reddit or videos on YouTube. But it can have more insidious effects too. “Under some conditions, even a minority opinion can appear to be extremely popular locally,” say Lerman and co. That might explain how extreme views can sometimes spread so easily. It might also explain the spread of antisocial behavior. Various studies have shown that teenagers consistently overestimate the amount of alcohol and drugs their friends consume. “If heavy drinkers also happen to be more popular, then people examining their friends’ drinking behavior will conclude that, on average, their friends drink more than they do,” say Lermann and co."