Developing data mining system using fuzzy association rules
![Thumbnail](/bitstream/handle/123456789/69778/karil%20revository%20ISS%204.pdf.jpg?sequence=4&isAllowed=y)
View/ Open
Date
2014Author
Sitanggang, Imas S.
Mustika, Arsha
Kustiyo, Aziz
Metadata
Show full item recordAbstract
Abstract. This research aims to develop 0. data mining system in order 10 extract association rules [rom 24%2 records 01 village potential data lor the \lear 2003 (PODES 2003) especiallv in the regions 01 Java. The algorithm used in this research named FIIZZYQuantitative Association Rules Mining is divided into three major parts including transforming the dara from the original format to jilzzy sets using the Fuzzy C-Means (FCM) algorithm. generating frequent itemsets. and extracting IIIZZY association rules. The results 01 this research SllOWS that 0. considerable number 01 rules have high fuu» confidence value because the value 01 [uu» support lor antecedents combined with their consequent are also high. The parameter that gives significant influence is minimum ji1zzy support (minsup). For minsup 90% and minimum luzzy confidence (minconf} 90%, the system generates 161uzzy association rules. For lift value 1.04, there are /IVO rules which show the relation 01 number oi family thai using electricity and number 01permanent building. Besides, lor mincorr value 0.8, there are [iv« rules whicn show the relation oi number ojunemplovment, number 01 students who dropped 01/1[rom elementarv school. number offamily that use electricity, and number ojpermanent building.
Collections
- Proceedings [2790]