Ant colony algorithm for graph based association rules mining

Ghassan Saleh Hussein Al-Dharhani (P47969)

Please use this identifier to cite or link to this item: https://ptsldigital.ukm.my/jspui/handle/123456789/476502

Full metadata record

DC Field	Value	Language
dc.contributor.advisor	Azuraliza Abu Bakar, Assoc. Prof. Dr.
dc.contributor.author	Ghassan Saleh Hussein Al-Dharhani (P47969)
dc.date.accessioned	2023-10-06T09:19:44Z	-
dc.date.available	2023-10-06T09:19:44Z	-
dc.date.issued	2011-06-17
dc.identifier.other	ukmvital:114685
dc.identifier.uri	https://ptsldigital.ukm.my/jspui/handle/123456789/476502	-
dc.description	A graph-based Association Rules Mining (ARM) is an approach to find frequent item sets from the transaction database in the graph representation. One of the graph based ARM algorithm is the Ant Colony Optimization (ACO) algorithm. The ACO is used to search the graph for the optimized path and generates the item sets as well as the association rules. Two challenges are encountered in the ACO algorithm for the graph based ARM that are i) how the data representation leads to the appropriate graph construction from the transaction database and, ii) how this data representation is integrated with the frequent items generation algorithm specifically the Apriori algorithm to speed up the execution time. Data representation is an essential stage for linked nodes in a weighted graph consisting set of vertices and edges. The aim of this study is to propose a data representation scheme thus improves the existing ACO algorithm for mining association rules (ACOMAR). The improvement of ACOMAR contains two phases. Firstly, a new transactional data representation scheme tailored for graph based ARM is proposed, and secondly, the improvement of the ACOMAR algorithm is made based on the proposed graph representation. The data representation involves several preprocessing steps that produce the customized Boolean Matrix representation. A standard Apriori Algorithm is applied to the represented data and the n-frequent item sets are generated. The improved ACOMAR relies on the graph from the 2-frequent items and uses the graph to generate the final frequent item set. The improved ACOMAR deals with simpler graph representation since the graph is generated from the 2-frequent itemsets instead of n-frequent item sets. Lastly, the final frequent item sets are generated from the graph by using the ACO algorithm. The simplicity of the ACOMAR is contributed by factors such as structure of the data, AVL-Tree and data representation of the transactions database. The experiments are conducted on the benchmark datasets by evaluating the performance of the improved ACOMAR towards the original ACOMAR and the apriori in terms of time, number of rules and the confidence of the rules. The experiment showed that improved ACOMAR generates the association rules in lesser time with comparable quality of rules.,“Certification of Master’s/Doctoral Thesis” is not available,Master Information Technology
dc.language.iso	eng
dc.publisher	UKM, Bangi
dc.relation	Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat
dc.rights	UKM
dc.subject	Ant algorithms
dc.subject	Mathematical optimization
dc.subject	Ants -- Behavior -- Mathematical models
dc.subject	Algorithms
dc.subject	Universiti Kebangsaan Malaysia -- Dissertations
dc.subject	Dissertations, Academic -- Malaysia
dc.title	Ant colony algorithm for graph based association rules mining
dc.type	theses
dc.format.pages	122
dc.identifier.callno	QA402.5.D495 2011 3 tesis
dc.identifier.barcode	002392(2011)
Appears in Collections:	Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat

Files in This Item:

File	Description	Size	Format
ukmvital_114685+SOURCE1+SOURCE1.0.PDF Restricted Access		11.64 MB	Adobe PDF	View/Open

Show simple item record Recommend this item