Forum Moderators: Robert Charlton & goodroi
...if it did would the "trust" not be based entirely on the who/how/why the site links OUT?
This kind of sincerity would be too easy to fake. Inbound links have got to be an important component.
PS - My hunch is that every little beta search tool that Google has out there is in some way being mined for "trust" data... and that Google is constantly looking for areas of agreement among these.
[edited by: Robert_Charlton at 11:31 pm (utc) on Feb. 17, 2007]
I've wondered about this quite a bit lately. I've even wondered if being "trusted" simply means an absence of "untrustworthy" signals, versus a site actually having signals of trust.
TrustRank White Paper [dbpubs.stanford.edu]- "Combating Web Spam with TrustRank"
Historical Data [webmasterworld.com] - WebmasterWorld thread about Google's 2005 patent [appft1.uspto.gov]
I'm pretty sure Google is looking for some positive signals, not just the absence of negatives.
As software started churning out pages and sites to fool the algo, other layers of trust measurement... positive and negative... have been added as factors in the algo.
If the question is about whether algos like Hilltop and Trustrank are really in use, look at the Google medical results seeded by experts. Can anyone imagine that Google isn't learning from this, and applying what's come out of Google Co-Op to the general results?
What's the whole class-C interlinking thing, if not about trustworthiness of link sources?
I'm assuming that there's every sort of mashup and statistical profiling imaginable, using data from wherever they can get it, to determine trust and reliability of sites. For me, the big question is what's practical for them at this stage of infrastructure.
Fair enough, though IBLs from similar topic pages may well get my CONTENT seen as "trusted", if those pages had off topic footer links I'd have my doubts G would see my SITE as "trusted".
That "Trustrank" thing seems to be a Y! rather than a G initiative, and reminds me very much of Gs' Localrank...
This thread on "Lots of trusted sites ranking on Internals alone" [webmasterworld.com...] got me wondering about "trust", if my OBLs fail to pass muster it might well have a knock on effect on my internals...
I'm assuming that there's every sort of mashup and statistical profiling imaginable, using data from wherever they can get it, to determine trust and reliability of sites. For me, the big question is what's practical for them at this stage of infrastructure.
Absolutely! - and if you invert this key statement, what's practical for a webmaster/siteowner to respond to?
Automated processes of mass reciprocal links, interlinked sites, deployment of syndicated affiliate content are now much less effective. Building trust takes time and great effort on the back of establishing soundly recognised referal relationships.
[edited by: Whitey at 10:36 pm (utc) on Feb. 18, 2007]
via the Google testing blog: [googletesting.blogspot.com...]
"Googlers in our Test Engineering group often speak at, and write for, forums on testing all over the world.of the logic applied to verifying the quality of results produced, of which "trusted results" formed a part, as seen here.
[stickyminds.com...]
We can—and generally need to—combine the various pattern-matching techniques to determine whether the factor is present and "correct."The scoring should be simple to code: If a pattern is matched correctly, return the respective score; otherwise return to zero. For cases such as the presence or absence of Web forms (such as bookNow), return the relevant score to +2 if found, otherwise -1. More complex scores can be implemented using case statements, etc. The value of some results will affect whether or not you evaluate other results. For instance if you detect an "HTTP 500" error you don't need to check the flight data on that page as the server has already reported an error.
By combining heuristic test oracles with the weighting of responses we can make our automated tests more powerful and reliable, even in volatile environments with complex dynamic data. The concepts are fairly simple, and you can work with experienced users to determine the factors and weightings. I hope you find these techniques useful. Let me know how you get on with these suggestions. Your comments and feedback will help to improve the material further.
It might provide some insight and translation, into how the engineers think alongside the algo's view of "validating" the trust of a result.
[edited by: Whitey at 4:14 am (utc) on Feb. 19, 2007]
