A Comparative Analysis of Tertius, Apriori, and FP-Growth Algorithm in Groceries Dataset

Abstract

A technique that can help search for patterns or associative relationships is the association rules technique. Association rules that are widely used are based on those that meet the minimum support and confidence requirements, these association rules are in the form of if antecedent then consequent. The association rule algorithm used is the Tertius algorithm, the Apriori algorithm and the FP-Growth algorithm, preceded by the preparation of a wholesale transaction database and the determination of the minimum support and confidence limits with a minimum support of 2%. This study aims to perform a comparative analysis of the Apriori algorithm and the FP-Growth algorithm to find purchasing patterns, using a wholesale dataset sourced from kaggle.com with a total of 4119 data, which is data from cafeteria wholesale sales. The results of the rules from the Apriori algorithm, the Tertius algorithm and the FP-Growth algorithm using the same support and confidence, obtained the same results. From transactions at Grocier with a minimum support of 30% and a minimum confidence of 60% are Rolls/Buns products with Pop plants. The conclusion is that if the buyer buys Rolls / Buns, there is a 75% chance of buying Pip fruit and Whole milk.