Talk:Affinity analysis

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Affinity analysis[edit]

Why'd you revert the line that I added to the Affinity analysis article? I've been trying to improve it to give a sense of how it's actually viewed and used. The report from AMR Research is behind a subscription wall, but they're a fairly respected market research group and I thought it gave a good perspective on what people were actually expecting from it, since much of the current article only talks about hypothetical situations and a few companies currently using it. If you have a better way to structure the article, go for it, but it's completely unhelpful to just revert good-faith edits. modargo (talk) 00:01, 31 December 2009 (UTC)[reply]

Website required login and is spam in my view. History2007 (talk) 00:07, 31 December 2009 (UTC)[reply]

Per WP:CITATION, requiring a login to view does not invalidate a reference if it's from an established organization, which AMR certainly is. If you want, I can quote the relevant section of the report on the article's talk page. modargo (talk) 00:19, 31 December 2009 (UTC)[reply]

How about quoting from the "thinker's guide to commonsense". What does a user gain from a website they can not access, about a report they can not read? How do you spell spam? That sentence told me NOTHING. Absolutely NOTHING except alert me to the fact that AMR has a report I may want to buy. How do you spell spam? If you do not like my view, ask for arbitration. I am not changing my mind on this. History2007 (talk) 00:30, 31 December 2009 (UTC)[reply]

Association algorithms[edit]

I think Association rule learning, K-optimal pattern discovery, etc. are specific instances of algorithms that implement Affinity analysis and one may consider merging them into here, with "Affinity analysis" as a general approach with no specific algorithms and the others as algorithmic methods of realizing it. But then it will make it a hard to read page. It would probably be better to refer to those algorithms from here, and keep the top level concept separate. History2007 (talk) 00:29, 3 January 2011 (UTC)[reply]

I agree with you regarding this organizational scheme. Next steps should then be to better link the seperate pages to one another. Jeffreydiehl (talk) 19:22, 3 January 2011 (UTC)[reply]
Ok, In the next few days I will come up with a link-scheme. A few of those need pages built, and I will put those on my to do list as well. Cheers. History2007 (talk) 19:36, 3 January 2011 (UTC)[reply]