GitHub user lxmly opened a pull request:
https://github.com/apache/spark/pull/18337
Fix batch gradient bug in SVDPlusPlus
## What changes were proposed in this pull request?
I think the mode of batch gradient in Graphx SVDPlusPlus is incorrect. The
changes of this commit are as blow:
* Use the mean of batch gradient instead of batch gradient.
* When calculating the bias of each rating during iteration, the bias
shouldn't be limit in (minVal, maxVal).
* The method "calculateIterationRMSE" is added to show the iteration effect
clearly.
## How was this patch tested?
I test SVDPlusPlus locally by using movielens-100k dataset. The RMSE of
origin SVDPlusPlus can't get converged and looks completely wrong. The RMSE of
new SVDPlusPlus gets converged correctly.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/lxmly/spark
svdplusplus_incorrect_batch_gradient
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/spark/pull/18337.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #18337
----
commit 7ed91593fb3056ae1e7df9962be06ccc2ce5f27c
Author: lxmly <[email protected]>
Date: 2017-06-17T10:28:15Z
fix svdplusplus batch gradient bug
----
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
