mahout-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Grant Ingersoll (JIRA)" <>
Subject [jira] [Updated] (MAHOUT-688) High Document Frequency pruning for seq2sparse
Date Fri, 06 May 2011 14:35:03 GMT


Grant Ingersoll updated MAHOUT-688:

    Attachment: MAHOUT-688.patch

Reorgs the code a little bit to move std. dev. calculation to a common place.  Also, puts
in tests for std. dev. and fixes the std. dev. calculation, which _I am pretty sure_ was incorrectly
calculated (it was missing the subtraction of the average in the sum of squares calc). Also
added license headers and cleaned up the formatting a bit.

Since we are in code freeze, we can iterate on this a bit for 0.6.

> High Document Frequency pruning for seq2sparse
> ----------------------------------------------
>                 Key: MAHOUT-688
>                 URL:
>             Project: Mahout
>          Issue Type: Improvement
>            Reporter: Vasil Vasilev
>            Assignee: Grant Ingersoll
>            Priority: Minor
>              Labels: Vectorization
>             Fix For: 0.6
>         Attachments: MAHOUT-688.patch, MAHOUT-688.patch
> This improvement allows to prune the words with high document frequencies from the tf
and tf-idf vectors produced by seq2sparse, based on the standard deviation of the words' document
frequencies and specifying which rods to be pruned in a means of times this standard deviation.
One good option is 3 times the standard deviation

This message is automatically generated by JIRA.
For more information on JIRA, see:

View raw message