An Enhancement of the Equivalence Class Transformation Algorithm Applied in Market Basket Analysis

ABSTRACT: Data mining has a major concern on precise specification on items in large databases. The equivalence class transformation algorithm is specifically applied in market basket analysis. Market Basket Analysis is used to know the statistics of items that are less salable and frequently sold. However, the researchers found some problems and limitations in the existing process. Through simulation of the existing algorithm the researchers found out that for having bigger Transaction ID set the computation time is costly when interesting. Along with this, due to its downward closure property, the infrequent items set often found on the later part of candidate generation. Lastly, through the use of uniform user-defined support count the number if item sets to be prune is numerous. Due to the slow performance in analysing huge volumes of data, the researchers came up with simplified process to a faster and more convenient transaction that will benefit both the consumers and dealers. The researchers introduce an enhancement to the existing process of the equivalence class transformation wherein it lessens the number of transaction to be process. The infrequent items can be easily distinguished and the upgrade of the use of support count is implemented. The enhancement of equivalence class transformation algorithm presented by the researchers is beneficial for easy, fast, and convenient transaction of items that will help both consumers and dealers.

5

ABSTRACT: Data mining has a major concern on precise specification on items in large databases. The equivalence class transformation algorithm is specifically applied in market basket analysis. Market Basket Analysis is used to know the statistics of items that are less salable and frequently sold. However, the researchers found some problems and limitations in the existing process. Through simulation of the existing algorithm the researchers found out that for having bigger Transaction ID set the computation time is costly when interesting. Along with this, due to its downward closure property, the infrequent items set often found on the later part of candidate generation. Lastly, through the use of uniform user-defined support count the number if item sets to be prune is numerous. Due to the slow performance in analysing huge volumes of data, the researchers came up with simplified process to a faster and more convenient transaction that will benefit both the consumers and dealers. The researchers introduce an enhancement to the existing process of the equivalence class transformation wherein it lessens the number of transaction to be process. The infrequent items can be easily distinguished and the upgrade of the use of support count is implemented. The enhancement of equivalence class transformation algorithm presented by the researchers is beneficial for easy, fast, and convenient transaction of items that will help both consumers and dealers.

Filipiniana

5