[ 
https://issues.apache.org/jira/browse/MADLIB-1028?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15571064#comment-15571064
 ] 

Enzo Mar commented on MADLIB-1028:
----------------------------------

I think a good use case for madlib could be for clickstream analysis: too much 
data to keep in memory.
Hence it is a good idea to beef up all functionalities around clickstream 
analysis (e.g. I'm going to report a bug on apriori related to this).
I haven't used the Aster functionality directly, but it seems interesting to be 
able to spot automatic clicking etc.

As a side note, to do a sessionization in sql, thanks to window functions, is 
not that difficult (and there are examples on the web).  Starting to add 
features like rapid_fire is more complex...

> add rapid_fire to sessionization
> --------------------------------
>
>                 Key: MADLIB-1028
>                 URL: https://issues.apache.org/jira/browse/MADLIB-1028
>             Project: Apache MADlib
>          Issue Type: Improvement
>          Components: Module: Utilities
>            Reporter: Enzo Mar
>            Priority: Minor
>
> The recent implementation of sessionization could benefit from adding a 
> "rapid_fire" field (a sort of min_filed in comparison to the existing 
> max_field).
> Aster Data implementation of sessionization has this field.
> Basically it is used to flag and potentially exclude events too close to each 
> other. This helps to detect / eliminate either errors or robotic type of 
> access.
> To see it in action have a look at the Aster video on sessionize: 
> https://www.youtube.com/watch?v=C760M9ttK9Q



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to