I love privacy invasions! (Data analysis)

danielwang · Post by **danielwang** » Fri Dec 12, 2003 11:37 pm

Suggestion only...
Not likely, but we could do something close, maybe...

Start tracking downloads of anime music videos and pageviews.
Link this data to the anime music video itself.

By treating each data property as a distinguished element in a group, we can use discrete group mapping to analyse the relationships between certain characteristics. That said, it works without any previous data on group bindings, it extrapolates the data statistically. In simple terms:

Keep track of the music videos that one person downloads. Then look at the other people who have downloaded that video. See what they've downloaded, and apply 4096 points to the analysis metric of that video. Continue down a few generations applying fewer points as you go along.
Map the videos to the music, indeed, keep track of the music in the videos that people download. Link that to videos groups, and to users.

There. statistical data on what people like.

And what is this algorithm good for?

---------------------

AnimeMusicVideos.org is NOT affiliated with Amazon.com

Download

You're about to download AniMix V (series: AniMix)

Other users also liked

Koopiskeva - Euphoria
"The highly acclaimed Euphoria. Deep and thematic, yet entertaining, Euphoria portrays"..... Read more
05m55s 4.78/5 8.65/10 (511 stars, 102 reviews) Reviews Why was I recommended this?

Kevin Caldwell - Caffeine Enconium
"Fun and lively, Caldwell's 'Caffeine Enconium' is worthy of it's name, and"..... Read more
04m17s 3.99/5 7.31/10 (78 stars, 32 reviews) Reviews Why was I recommended this?

Free Shipping
All downloads at the Org are absolutely free. You're welcome to donate!

angelx03 · Post by **angelx03** » Fri Dec 12, 2003 11:47 pm

Declan_Vee · Post by **Declan_Vee** » Sat Dec 13, 2003 12:05 am

Whoa... Umm, all I can say is. Isn't that a bit excessive?

danielwang · Post by **danielwang** » Sat Dec 13, 2003 12:10 am

Declan_Vee wrote:Whoa... Umm, all I can say is. Isn't that a bit excessive?

Not nearly as excessive as, say, a specific, certain, online retailer...

You know who I mean.

dokidoki · Post by **dokidoki** » Sat Dec 13, 2003 1:04 am

Um, there are already plans for something like that...

Scintilla · Post by **Scintilla** » Sat Dec 13, 2003 3:32 am

dokidoki wrote:Um, there are already plans for something like that...

But no one's responded to that thread since June... is anything actually happening on that front, or did the idea get forgotten?

trythil · Post by **trythil** » Sat Dec 13, 2003 4:10 am

Scintilla wrote:
dokidoki wrote:Um, there are already plans for something like that...
But no one's responded to that thread since June... is anything actually happening on that front, or did the idea get forgotten?

It's more of a "there are more important things to do" kind of thing.

downwithpants · Post by **downwithpants** » Sat Dec 13, 2003 2:19 pm

i like wang's idea - having recommendations on individual videos. the problem is videos like euphoria, tainted donuts, and odorikuruu would show up on every recommendation, and none of the lesser known videos would never show up.

Lyrs · Post by **Lyrs** » Sat Dec 13, 2003 2:57 pm

danielwang wrote:Start tracking downloads of anime music videos and pageviews.Link this data to the anime music video itself.

This reminds of a thread i made a month or so back:
http://lennertmods.sourceforge.net/mods/bbus/

lol, Phade is watching your every move.

danielwang · Post by **danielwang** » Sat Dec 13, 2003 3:00 pm

downwithpants wrote:i like wang's idea - having recommendations on individual videos. the problem is videos like euphoria, tainted donuts, and odorikuruu would show up on every recommendation, and none of the lesser known videos would never show up.

The system could be "fixed" RIGGED XD to show some of the lesser known videos. THat way, we could make sure that all of the good videos get attention... and save people from the good ones that they don't really want.

A rotational turnover would ensure that people aren't "fixed" into one interest group by continually following the links, we'll call it "Discover New AMVs" or such, and that AMVs on the top aren't held there by continuous pageviews.

The algorithm, although it sounds complicated, is probably trivial for somebody like Phade. And since we wouldn't disclose user preferences, there wouldn't be much objection, no?

I love privacy invasions! (Data analysis)

I love privacy invasions! (Data analysis)

Re: I love privacy invasions! (Data analysis)