You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@mahout.apache.org by VIGNESH PRAJAPATI <vi...@gmail.com> on 2011/10/18 23:40:10 UTC
mahout Algorithm for recommendation engine
Hello, mahot users.
I am vignesh prajapati from india want to develop recommendation engine with
hadoop , mahout and hive.
But after using
http://code.google.com/p/unresyst/wiki/CreateMahoutRecommender information i
have generated recommendations via slopbased algorithm.And for generating
recommendations it is taking a 18 minutes and generating same
recommendations for all of users among 51,970 users and 14,000 items.
Ans after if i am normalizing all of data in csv files then its taking 4.5
seconds with no recommendations. so is there any solution and suggestion for
this ?
Thanks in advance..
--
*Vignesh Prajapati* ****
Tel: 9427415949 |
vignesh2066@gmail.com | www.makeyour.tk
MYTK [image: Facebook] <https://www.facebook.com/vigs143> [image:
Twitter]<http://twitter.com/#%21/vigs143> [image:
LinkedIn] <http://www.linkedin.com/pub/vignesh-prajapati/37/756/46a> [image:
about.me] <http://www.makeyour.tk/><http://r1.wisestamp.com/r/landing?promo=7&dest=http%3A%2F%2Fwww.wisestamp.com%2Femail-install%3Futm_source%3Dextension%26utm_medium%3Demail%26utm_campaign%3Dpromo_7>
Re: mahout Algorithm for recommendation engine
Posted by Sean Owen <sr...@gmail.com>.
Slope-one? This implementation has nothing to do with Hadoop or Hive.
What are you doing, exactly?
Generating 51,970 user recommendations in 18 minutes isn't bad. That's
20ms per user, though still sounds a bit slow for this data set size.
I don't know what "normalizing" means here. Something sounds wrong
with your input if you get all the same results.
On Tue, Oct 18, 2011 at 10:40 PM, VIGNESH PRAJAPATI
<vi...@gmail.com> wrote:
> Hello, mahot users.
>
> I am vignesh prajapati from india want to develop recommendation engine with
> hadoop , mahout and hive.
> But after using
> http://code.google.com/p/unresyst/wiki/CreateMahoutRecommender information i
> have generated recommendations via slopbased algorithm.And for generating
> recommendations it is taking a 18 minutes and generating same
> recommendations for all of users among 51,970 users and 14,000 items.
> Ans after if i am normalizing all of data in csv files then its taking 4.5
> seconds with no recommendations. so is there any solution and suggestion for
> this ?
>
> Thanks in advance..
>
> --
>
> *Vignesh Prajapati* ****
> Tel: 9427415949 |
> vignesh2066@gmail.com | www.makeyour.tk
> MYTK [image: Facebook] <https://www.facebook.com/vigs143> [image:
> Twitter]<http://twitter.com/#%21/vigs143> [image:
> LinkedIn] <http://www.linkedin.com/pub/vignesh-prajapati/37/756/46a> [image:
> about.me] <http://www.makeyour.tk/><http://r1.wisestamp.com/r/landing?promo=7&dest=http%3A%2F%2Fwww.wisestamp.com%2Femail-install%3Futm_source%3Dextension%26utm_medium%3Demail%26utm_campaign%3Dpromo_7>
>