Slashdot Mirror
Google to use TrustRank for News, Possibly More
mike slaven writes "In a follow-up to Tuesday's post about Google registering a trademark on the term TrustRank, an article on NewScientist explains how Google plans to track the credibility of news sources. The article also mentions that the patent on TrustRank is not limited to ranking just news stories: 'The patent also reveals that the same system could be roped in to rank other search results, not simply news. So sales and services could in the future be listed on the basis of price and the reputation of the company involved.'"
I'll be fascinated to see how Slashdot articles get ranked.
Now for the gamespy implimentation to see what servers are filled with lamers... Now that'd be useful!
Well, this is (half) a shameless plug but my newsbot has been ranking news sources and referring meta-news sources according to "trust" for over 3 yrs now. Findory (no affiliation) does something similar by ranking each individual news story.
From what I have read about trust rank, the basic premise is that they pick 200 or so "trusted sites". The trust rank for any page is then basically the number of link hops to the page from a trusted site.
I was gonna mod you down, but I'll post instead. If you'd have RTFA, you'd have seen that the above statement is totally false. There are lots of criteria they are using to generate the number, of which internet traffic is only 1. To quote TFA:
he database will be built by continually monitoring the number of stories from all news sources, along with average story length, number with bylines, and number of the bureaux cited, along with how long they have been in business. Google's database will also keep track of the number of staff a news source employs, the volume of internet traffic to its website and the number of countries accessing the site.
Google will take all these parameters, weight them according to formulae it is constructing, and distil them down to create a single value. This number will then be used to rank the results of any news search.
"The problem with internet quotations is that many are not genuine" -Abraham Lincoln
Given that ALL news sources or sites have some bias, isn't trust in the eye of the beholder in many cases? Although the objective facts of some situation may be undisputed (usually these are disputed, too), the interpretation of those facts is subjective and highly dependent on the viewpoint, world model, etc.
Perhaps Google will need to introduce right-wing and left-wing versions of TrustRank. If it does not, then it will be an example of tyranny of the majority when Google asserts than the majority's bias is trustworthy.
Two wrongs don't make a right, but three lefts do.
There are two main ways of ranking news (as well as websites and whatnot): Popularity or Quality. Popularity can be easily quantified, but with news you will probably end up with big and well known media source dominating the ranking. Regardless of what one thinks about CNN or BBC, they are not what I look for in googlenews (or I would just type www.cnn.com). I want news sources that I was anware of, to give me a fresh perspective.
Quality on the other hand is very hard to measure, and any definition will surely be controversial. The metrics they are proposing will also benefit large well knwown news sources.
What I would really like though is a rank that gave the widest possible perspective. Some algorithm that would take a news event and define some broad categories of news sources (say, on news on Iraq, conservative american media, liberal american, iraqi, arab, etc.). Then one or two representatives of each category would be displayed in the ranks, choosen among the cotegory by the metrics described in the article. The trick of course is to define the categories, but I think one could do that by looking at how different sources 'cluster' together. Sources in one country link overwhelmingly among themselves rather than abroad. Conservative news sources tend to cite and link to other conservatives, ditto for liberals or any other category. This is even more true for blogs, which wouldn't be much cited in the rank itself, but are a sort of glue that binds ideological and national categories together, and thus provide useful information to help classify the news source.
I hope I'm making sense here. Just my 2cents...
I got your prior art http://slashdot.org/comments.pl?sid=116588&cid=986 5259">right here.
How about over time Google learns which sites I trust and modifies the ranking according to that. We could give each site a thumbs up or down rating like you do on a TiVO.