Market Basket Analysis Using FP-Growth Algorithm

Abstract

Data mining is the mining or discovery of new information by searching for a particular pattern or rule of a very large amount of data, its utilization is already widely applied in various fields. One of the data mining techniques is Market Basket Analysis which is used to find associations between different product sets that customers place in a basket. This thesis will discuss Market Basket Analysis using the Frequent Pattern Growth (FP-Growth) algorithm on the lending data of books at the Library of Islamic University of Bandung. The purpose is to find associations among books that are often borrowed simultaneously measured through the support value which is a measure that shows how much chance of lending transactions containing the itemsets borrowed simultaneously. The confidence value is a measure that shows how much the association between two books borrowed simultaneously from all transactions containing one of the books and found strong association rules of elevated ratio values based on frequent items Obtained. There are 9,804 lending transactions comprising of 250 book codes. The results indicate that the books (codes 657 and 658), (code books 346 and 297) and (1001 and 302 book codes) have support, confidence, and lift ratio values that meet the minimum requirements. Then the three itemsets can be used as a recommendation for the borrowers of books visiting the Library of Islamic University of Bandung and the book can be placed side by side.