8

The documentation for the very powerful open source data mining tool SPMF lists them separately:

http://www.philippe-fournier-viger.com/spmf/index.php?link=algorithms.php

Does any one know why?

R Claven
  • 1,160
  • 2
  • 13
  • 27

3 Answers3

7

The web site has a pretty detailed Q&A flowchart which shows what the difference is from the author's point of view:

http://www.philippe-fournier-viger.com/spmf/map_algorithms_spmf_data_mining097.png

Tom Morris
  • 10,490
  • 32
  • 53
7

Glad that you are using SPMF. I wrote a post a while ago explaining the distinction between sequential patterns and sequential rules:

http://data-mining.philippe-fournier-viger.com/introduction-to-sequential-rule-mining/

If you read this blog post, the distinction will become clear. Basically, the main difference is that sequential patterns are only found on the basis of how frequent they are, while sequential rules also consider the probability of confidence that a pattern will be followed. Thus sequential rules are more useful for task such as doing predictions.

Phil
  • 3,375
  • 3
  • 30
  • 46
  • I am biased towards the usage of Sequential Rules Mining for mining applications involving sequences. And I mean, I am not able to differentiate between the applications of both of these mining ideas. Can you elaborate on the types of applications where a specific approach (spm or srm) to be used? – Ramvignesh Nov 07 '18 at 06:37
  • @Ramvignesh -- see this decision tree (http://www.philippe-fournier-viger.com/spmf/map_algorithms_spmf_data_mining097.png). On the left side it asks "Do you want patterns having confidence or probability (for example to perform predictions)?" If Yes -> Sequential Rules; If No -> Sequential Patterns. – Ian Thompson Mar 25 '22 at 17:52
1

Association rules are only one kind of patterns.

Most likely these algorithms can only find this kind of patterns because they employ some.interestingness measure to prune candidates?

Has QUIT--Anony-Mousse
  • 76,138
  • 12
  • 138
  • 194