Q-VIPER: Quantitative Vertical Bitwise Algorithm to Mine Frequent Patterns

Czubryt, Thomas J.; Leung, Carson K.; Pazdor, Adam G. M.

doi:10.1007/978-3-031-12670-3_19

Cited by 9 publications

(1 citation statement)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Furthermore, the collection of the useful rule gets the attention of the user, but not the whole rules related to the huge amount of database led to the mining process for association rules depending on metaheuristic algorithms. If we focus on the classical algorithms, we will notice that they search on the high-quality rules but not all the intact rules 3,4 .…”

Section: Introductionmentioning

confidence: 99%

خوارزمية فصيلة ميركات المعدلة لتعدين القواعد الرابطة

Saleh,

Sadiq

2024

Baghdad Sci.J

View full text Add to dashboard Cite

Association Rules Mining (ARM) forms one of the important data mining techniques. The classical methods that were previously worked on by researchers have become ineffective to deal with the steady growth of databases, which prompted us to use the mining process for association rules based on metahuristic, and in our work all the correct rules will extracted, and mining is not limited to high-quality rules. Swarm intelligence based is one of these methods. In this paper, Modified Meerkat Clan for Association Rules Mining (MCC-ARM) has been proposed. Basically, the proposed algorithm depends on Meerkat Clan Algorithm (MCA). The greatest benefit is the diversity of candidate solutions in MCA. In our work the rules will represented using two methods which are borrowed from the genetic algorithm; in the first one each group of rules refers to object in society which is called Pittsburgh; while the second one each rule refers to an object in society which is called Michigan. The proposed algorithm aims to inspect for the maximum possible number of correct association rules. The so-called algorithm follows the approach of defining the effective search area, which depends on a main random mechanism to lead the algorithm in extracting alternative rules and avoiding total solutions from being guided by the same rule, and this led to a great deal of diversity. In addition, the MCC-ARM uses condensation method in the adjacency search process to prevent the algorithm from falling into the local mode. In order to prove their efficiency, it should be applied on four reliable datasets (i.e. Zoo, German Credit, Primary Tumor and Chess). The enhancement brought about by the proposed algorithm has obtained two crucial factors, namely on the number of correct rules and quality fitness value.

show abstract

Section: Introductionmentioning

confidence: 99%