[
https://issues.apache.org/jira/browse/MAHOUT-157?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12763929#action_12763929
]
Isabel Drost commented on MAHOUT-157:
-------------------------------------
Great work Robin. I just had a look at the code and only found some minor
things:
ParallelFPGrowth
- it might be a good idea to reuse the DefaultOptionCreator to generate common
options like input and output.
- I would love to see a help option as well.
- What happens, if the users gives the wrong parameters? As a user, I would
rather not get confronted with a stack trace, even though it is an example.
- did you provide details on how to run the algorithm, the assumptions it
makes, file format, behaviour if the output file exists already on the wiki?
- the class is named ParallelFPGrowth, but if I read it correctly, it looks
like the entry point for both, the parallel and sequential version. Maybe
rename to FPGrowthJob?
FPGrowth
- line 98 is this really a recoverable error that does not cause
inconsistancies later on? Log message says "this should not happen" - what if
against all odds, it does happen? Why not throw a non-Checked Exception?
- line 177 we should not have source code that is commented out in newly added
code.
- The class seems to implement both - top k and vanilla fp growth - would it
make sense to split that up into different classes?
- generateFrequentPatterns - maybe it is just me, but I am always happy to find
tiny little comments in methods that long that very shortly explain what the
following code block is doing.
FPTreeDepthCache
- maybe mention in the docs that the implementation is not threadsafe?
FPTree, Pattern
- missing a class comment.
Pattern
- line 173 - please remove code that is commented out
AggregatorMapper
- the reporter is left unused.
Nice-To-Have: It would be nice to have package level comments in JavaDoc as
well.
> Frequent Pattern Mining using Parallel FP-Growth
> ------------------------------------------------
>
> Key: MAHOUT-157
> URL: https://issues.apache.org/jira/browse/MAHOUT-157
> Project: Mahout
> Issue Type: New Feature
> Components: Frequent Itemset/Association Rule Mining
> Affects Versions: 0.2
> Reporter: Robin Anil
> Assignee: Robin Anil
> Fix For: 0.2
>
> Attachments: MAHOUT-157-August-17.patch, MAHOUT-157-August-24.patch,
> MAHOUT-157-August-31.patch, MAHOUT-157-August-6.patch,
> MAHOUT-157-Combinations-BSD-License.patch,
> MAHOUT-157-Combinations-BSD-License.patch,
> MAHOUT-157-inProgress-August-5.patch, MAHOUT-157-Oct-1.patch,
> MAHOUT-157-Oct-8.pfpgrowth.patch,
> MAHOUT-157-Oct-8.TestedMapReducePipeline.patch,
> MAHOUT-157-September-10.patch, MAHOUT-157-September-18.patch,
> MAHOUT-157-September-5.patch
>
>
> Implement: http://infolab.stanford.edu/~echang/recsys08-69.pdf
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
