The Data Mining Forum 
IMPORTANT: This is the old Data Mining forum.
I keep it online so that you can read the old messages.
Please post your new messages in the new forum: https://forum2.philippe-fournier-viger.com/index.php
Sequence Analysis
Posted by:
Bob Camilleri
Date: January 14, 2015 02:21PM
I hope this group can direct me towards some resources or research. I am trying to build some models of a web user's behavior. In some cases, the order in which a user performs actions makes a difference, i.e. A -> B -> C is not the same as B -> A -> C. My aim is to encode sequences of actions (very very long sequences) and then to cluster users by these sequences.
The variety of actions possible is also very large, so probably it requires some generalisation at this stage. Any suggestions on what methods to look at?
The variety of actions possible is also very large, so probably it requires some generalisation at this stage. Any suggestions on what methods to look at?
Re: Sequence Analysis
Posted by:
webmasterphilfv
Date: January 15, 2015 09:58AM
You could search for papers related to sequence clustering by adding keywords such as "web" and "clicks"
By searching quickly, I have found that paper in KDD 2003:
http://web.itu.edu.tr/~sgunduz/papers/kdd.pdf
It seems to perform clustering of sequences of clicks on webpages.
Then, you could also look at recent work that have cited that paper as a starting point.
Otherwise, there may be some other newer papers on this topic.
Besides, I know that many papers have been done on sequence clustering but for biological data. I'm not sure if these work may be transfered to your problem but if you look at some survey on these work perhaps you could maybe find some ideas from that too.
By searching quickly, I have found that paper in KDD 2003:
http://web.itu.edu.tr/~sgunduz/papers/kdd.pdf
It seems to perform clustering of sequences of clicks on webpages.
Then, you could also look at recent work that have cited that paper as a starting point.
Otherwise, there may be some other newer papers on this topic.
Besides, I know that many papers have been done on sequence clustering but for biological data. I'm not sure if these work may be transfered to your problem but if you look at some survey on these work perhaps you could maybe find some ideas from that too.