Re: "Similar Charts" Still Needs That Overhaul

albummaster — Mon, 21 Sep 2015 05:11:17 GMT

Author: albummaster

Posted: 09/21/2015 09:11

Agree, there should be some weighting so that less common selections have more weight. but at the moment all entries are treated equally. The similarity is currently based on the number of albums and artists in common with other charts.

"Similar Charts" Still Needs That Overhaul

Applerill — Sun, 13 Sep 2015 19:42:48 GMT

Author: Applerill

Posted: 09/13/2015 23:42

I posted about this years ago, but I really don't feel any differently about it today with my broader taste.

Why does the "Similar Charts" function point me to users like TheHolyKiwi and BadSeed (both good users, mind you) instead of my real influences like Precedent, Permafrost and DBZ? The above users are great, but I have next to nothing in common with them in regards to music taste.

The obvious reason for this happening is that the Similar Charts algorithm just picks up a few canon albums that everyone has on their charts anyway, instead of weighting them by saturation like they should. In a perfect world, all my pop and experimental albums would be enough to put me with either a hardcore poptimist or an Abstract-Qualities Poeseur.

I just moved a bunch of canon stuff off my overall to test (with lesser-known albums I love just as much taking their place, don't worry), and even with only 9 albums in common, the same thing happens.

What do you guys think? Should we try to make the algorithm for Similar Charts more dynamic?

Best Ever Albums

Re: "Similar Charts" Still Needs That Overhaul

"Similar Charts" Still Needs That Overhaul