IMPLEMENTASI ALGORITMA APRIORI UNTUK MENEMUKAN FREQUENT ITEMSET DALAM KERANJANG BELANJA

Abstract

Apriori algorithm use an iterative approach where k-itemset used to explore (k+1)-itemset. (K+1)-itemset candidates containing subset frequency that rarely appears will not used in determining association rules. Association rules formed by “if antecedent then consequent”. Implementation of the apriori algorithm was preceded by the preparation of transactions database and determination of minimum support and confidence. Apriori algorithm scanning database repeated, pair one item to another and record the number of occurrences in the overall transaction. Frequent itemset is determined by selecting a combination or itemset that the count value greater than or equal to the minimum support then calculated the percentage value of support and confidence of each candidate. The association rules selected from which fit the minimum support and confidence. Data used in this study was sample of 100 transactions from database point if sales. Final Association rules are obtained from implementation of apriori algorithm is "if HELLO PANDA REFIL then HELLO PANDA 10gr" with percentage of support 2.00% and confidence 100.00% and "if HELLO PANDA 10gr then HELLO PANDA REFIL" with percentage of support 2.00% and confidence 100.00%. So it can be concluded that most customers buy HELLO PANDA REFIL will also buy HELLO PANDA 10gr so goes otherwise. This study proves that the apriori algorithm suitable implemented to search the frequent itemset in the shopping cart. Association rules that formed of frequent itemset can be used as decision support in sales.