API for custom actions and defining weights

[thelinuxlich]

So we can be more expressive and tailor the reccomendation for every need

[guymorita]

Can you give some examples? What's the use case? How would you like the
interface to look?
On Mon, Nov 24, 2014 at 1:00 PM Alisson Cavalcante Agiani <
[email protected]> wrote:

So we can be more expressive and tailor the reccomendation for every need

—
Reply to this email directly or view it on GitHub
#23.

[joernroeder]

I just sumbled over this issue while i'm researching if it's possible to weigth a specific event like so:

raccoon.liked('garyId', 'movieId1', 0.1)
raccoon.liked('garyId', 'movieId2', 0.9)

I saw that you're using redis sadd under the hood, therefore implementing weights would require some refactoring in the input and the algo as well. Weighted events would be a nice feature and would allow to add recurring events (like views) to the engine. what do you thing?

[guymorita]

Hi Jorn,

Thanks for reaching out. It's definitely possible. It would require a fair
amount of refactoring. Want to take a stab?

Can you walk me through the use case? Thanks!

On Wed, Aug 26, 2015 at 2:25 PM, Jörn Röder [email protected]
wrote:

I just sumbled over this issue while i'm researching if it's possible to
weigth a specific event like so:

raccoon.liked('garyId', 'movieId1', 0.1)
raccoon.liked('garyId', 'movieId2', 0.9)

I saw that you're using redis sadd
https://github.com/guymorita/recommendationRaccoon/blob/master/lib/input.js#L33
under the hood, therefore implementing weights would require some
refactoring in the algo as well. Weighted events would be a nice feature
and would allow to add recurring events (like views) to the engine. what do
you thing?

—
Reply to this email directly or view it on GitHub
#23 (comment)
.

Guy Morita

T 206.240.5846 l [email protected] [email protected]

LinkedIn http://www.linkedin.com/in/guymorita l Github
https://github.com/guymorita l Developer Blog
http://guymorita.tumblr.com l Svbtle http://alchemist.svbtle.com

[joernroeder]

hey, i was bussy doing other things but i had some spare time this week to look back into this old project.

a possible use case would be a user visiting a page multiple times and summing up the weights.

raccoon.liked('garyId', 'movieId1', 0.1)
raccoon.liked('garyId', 'movieId1', 0.1)

In this case garyId liked movieId1 with a score of 0.2. Another feature would be a custom function to increment/calulate the score.