__/ [ Big Bill ] on Monday 01 May 2006 21:42 \__
> On Mon, 01 May 2006 18:10:07 +0100, Roy Schestowitz
> <newsgroups@xxxxxxxxxxxxxxx> wrote:
>
>>__/ [ Big Bill ] on Saturday 29 April 2006 20:19 \__
>>> What's the way that Amazon determine the "like that" choices, you
>>> know, people who bought that book also bought, etc? What's the proper
>>> name for it?
>>
>>
>>I think they are using users' profile/history to study correlation between
>>books. They have a large sample set to learn from. It's a bit like spam
>>filters, only in a positive setting.
>>
>>
>>It's probably somewhat of a contextual reference. I am not sure what they
>>call it, but they patented the idea, which I think is outrageous.
>
> They apparently share it, there's an online radio station using
> RealAudio that uses it. Not Pandora, although that's arguably similar,
> something else.
It's not unprecedented. Some music player that I use (AmaroK) uses similar
ideas. I wrote about it last year and even included a reference to a radio
station that did exactly that.
http://schestowitz.com/Weblog/archives/2005/07/06/playlist-similarity/
The method has become rather outspread recently. iTunes is no exception as it
recommends music online.
>>I opine
>>that since the very dawn of time people knew that people's past preferences
>>reflected on future preferences and friends' choices could be treated as
>>though they were recommendations. Amazon's patent is the electronic
>>equivalent, only phrased more eloquently and technically in a patent
>>application, which I believe was filed and authorised. Courts will tell if
>>its value holds.
>
> Ta.
Best wishes,
Roy
--
Roy S. Schestowitz | #FFFFFFF4 ADD &R1, "9999999", &BankAccount
http://Schestowitz.com | SuSE Linux ¦ PGP-Key: 0x74572E8E
5:00am up 4 days 11:57, 13 users, load average: 0.38, 0.59, 0.71
http://iuron.com - help build a non-profit search engine
|
|