[
https://issues.apache.org/jira/browse/SPARK-13103?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15130964#comment-15130964
]
holdenk commented on SPARK-13103:
---------------------------------
I think that is expected behavior - if your finding too large a number of
collisions you can change the number of features with setNumFeatures
> HashTF dosn't count TF correctly
> --------------------------------
>
> Key: SPARK-13103
> URL: https://issues.apache.org/jira/browse/SPARK-13103
> Project: Spark
> Issue Type: Bug
> Components: MLlib
> Affects Versions: 1.6.0
> Environment: Ubuntu 14.04
> Python 3.4.3
> Reporter: Louis Liu
>
> I wrote a Python program to calculate frequencies of n-gram sequences with
> HashTF.
> But it generate a strange output. It found more "一一下嗎" than "一一下".
> HashTF gets words' index with hash()
> But hashes of some Chinese words are negative.
> Ex:
> >>> hash('一一下嗎')
> -6433835193350070115
> >>> hash('一一下')
> -5938108283593463272
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
