<![CDATA[PT XYZ is a company that provides livestock production facilities. Sales transactions are recorded as company files, sales reports, and income statements. More than 1,500 invoices are printed every month. However, in terms of product promotion, the company have not used the analysis results from the history of sales transactions. This study aims to provide product recommendations using the ECLAT algorithm. The ECLAT (Equivalence Class Transformation) algorithm uses the concept of depth-first search to find itemsets that often appear in transactions. The research steps are interviews for data acquisition, data pre-processing, data transformation, and data mining process with the ECLAT algorithm to find frequent itemsets and use the frequent itemset results as the basis for making association rules patterns. The results of the analysis show that the system can provide recommendations for association rules effectively from 14,617 transaction history. The highest minimum support that can be used to find a combination of k-itemset is 1%. The results of the annual association rules from the transaction history in 2018-2020 show the difference in results with the highest variance occurring in 2020, namely 5 association rules. Each association rule that appears has a strong confidence value that is above 50%]]>
Copyrights © 2021
