PENERAPAN ALGORITMA APRIORI UNTUK IDENTIFIKASI POLA KETERKAITAN PEMINJAMAN BUKU PADA PERPUSTAKAAN STMIK MULIA DARMA

Abstract

Libraries generate borrowing transaction data that can be utilized to identify relationships among books frequently borrowed in the same transaction. This study applies the Apriori algorithm to identify patterns of relationships among book-borrowing transactions at the STMIK Mulia Darma Library. The research stages include data selection, data cleaning, transformation into a transaction basket format, frequent itemset generation, association rule generation, and evaluation using support, confidence, and lift ratio. Since the available research files do not contain actual borrowing transaction data from the STMIK Mulia Darma Library, the quantitative results reported in this study use 360 anonymous/simulated test transactions to verify the analysis mechanism. Under the main scenario, with a minimum support of 5% and a minimum confidence of 70%, the algorithm generates 50 frequent itemsets and 39 association rules with a lift ratio greater than one. Testing six combinations of threshold values shows that increasing the minimum support and minimum confidence reduces the number of patterns, making the resulting rules more selective. The verification results indicate that the Apriori algorithm can be used to transform borrowing records into information about relationships among library collections, which can support shelf arrangement, reading recommendations, and collection acquisition planning. The numerical results must be replaced or validated using official data before being used as a basis for library decision-making.