Evaluation of rule interestingness measures in medical knowledge discovery in databases

An Efficient Association Rule Mining Me-thod for Personalized Recommendation in Mobile E-commerce

“…However, those algorithms remain powerless when dealing with the growing mass of business information. Especially when mining frequent k-Itemsets (k>2), BitTableFI and other Apriori liked algorithms are still with low efficiency; In the other respect, towards reducing the number of redundant frequent itemsets, some researches suggested using different interestingness measures for determination of association rules [10] , but they were not is suitable for personalized recommendation [11] .…”

“…Definition.7 Interestingness: For a random rule like R: A B in D, the interestingness of R is defined in Eq. (10), , I R P A B P A P B V R (10) where, the function V(R) is the validity [10] , I [-0.25,0.25]. For any rule like R: A B, its validity is the difference between two probabilities, shown in Eq.…”

MOCA-I: Discovering Rules and Guiding Decision Maker in the Context of Partial Classification in Large and Imbalanced Datasets

“…Over 38 common rule interestingness measures are referenced by Geng and Hamilton in their review [5], while Ohsaki et al studied measures used in medical domain [6] and Greco et al studied Bayesian confirmation measures [7]. Table 1.…”

A New Definition and Look at DNA Motif

“…For a review of the measures refer to (McGarry, 2005;Geng and Hamilton, 2006). Some researchers (Ohsaki et al, 2007) showed that these measures can fairly represent expert needs in a specific domain. It can be shown that deployment of some other measures will also lead to insignificance of TTG.…”