In Apriori Association Rule if the minSupport = 0.25 and minConfidence = 0.58 and for an item set we found a total of 16 association rules:
Rule Confidence Support
{1 2 ==>3} 1 0.4
{3 5 ==>2} 1 0.4
{1 ==> 2 3} 0.666 0.4
{1 3 ==> 2} 0.666 0.4
{2 3 ==> 1} 0.666 0.4
{5 ==> 2 3} 0.666 0.4
{2 3 ==> 5} 0.666 0.4
{2 5 ==> 3} 0.666 0.4
{1 ==> 3} 1 0.6
{5 ==> 2} 1 0.6
{3 ==> 1} 0.75 0.6
{2 ==> 3} 0.75 0.6
{3 ==> 2} 0.75 0.6
{2 ==> 5} 0.75 0.6
{5 ==> 3} 0.666 0.4
{1 ==>2} 0.666 0.4
If we want to reorder these rules from the most to least important rules which factor determine the importance of the rule Support or confidence i.e:
In this rule the Confidence is 1 but the Support is 0.4
{1 2 ==>3} 1 0.4
While in this rule the Confidence is 0.75 but the Support is 0.6
{3 ==> 1} 0.75 0.6
In the support-confidence framework (proposed in the early 1990's), the support of an association rule A==>B (percentage of transactions that contain both A and B) is employed to indicate the rule’s relevance, whereas the confidence index (the probability that a transaction contains B, given that it contains A) is used to indicate the validity of the association rule. Therefore, both interest measures are important.
Actually, in the majority of real-world association rule minining processes, confidence is used tandem with some other interest measures (such as lift, correlation coefficient, conviction, etc.) in order to determine the strength of an association rule. More information about the pros and cons of the confidence index can be found at the below paper:
Conference Paper Mining Strong Associations and Exceptions in the STULONG Data Set
- Badges
- Science topic
- Similar topics
- Computer Science
- Data Mining