Recommendation engines

Music recommendation engines have a long long way to go.

I never liked Pandora – repertoire was too US centric and the academic model for recommendation (e.g. you chose a synth rock song in a minor key with prominent guitars, therefore here’s more synth rock in a minor key) was lamentable.

I have been an occasional user of last.fm for a couple of years but it seems to perform worse with age. The logic of recommendation based on scrobbling what people actually listen to seemed like such a good idea but last.fm now returns such poorly worked out recommendations that I do not believe scrobbling is at the core of it.

It seems to stay in very poorly defined grooves. So if you picked a well known 80s pop band it will feed endless well known 80s pop band even if you are constantly skipping tracks.

Choose an unknown band and it only feeds you unknowns. No one listens to music in this way. That’s why scrobbling was a cool idea. Could it be that it doesn’t work as a way of making good recommendations? Certainly, in comparision, the certainty of a known album or the ‘editorial voice’ of a good radio show or dj mix beats last.fm.

Over 50% of tracks I skip within the first minute. What a waste of time and it makes it impossible to get into the flow of the music.

At this rate I will go back to listening to albums!

But not before I check out TheFilter.

Advertisements
Explore posts in the same categories: Music

Tags:

You can comment below, or link to this permanent URL from your own site.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: