HEllo,
I guess this is a rather simple question but I am new to the topic.
On order to create assoziation rules from sequence databases do you always have to mine the sequential pattern first? And if so, is there some literature on this?
thanks

Hi,

The answer is no.

There are some algorithm like Rulegen by Zaki (2001) who will first mine sequential patterns in a sequence database and then use them to create some sequential rules:

http://www.philippe-fournier-viger.com/spmf/Zaki2001.pdf

But some other algorithms like RuleGrowth and some others will directly find the rules:

http://www.philippe-fournier-viger.com/spmf/TKDE2015_sequential_rules.pdf

These two types of rules are a little bit different though. You can check the RuleGrowth paper in the introduction for a brief comparison discussion of that.

Best regards,

Philippe

Thank you very much for the quick response and thank you for all the work on SPMF. It helped me a lot.

My pleasure. Happy that SPMF is useful ;-) Thanks for using it :-)

Philippe