The FP-Growth and Fuzzy C-Covering algorithms are known to correct the Apriori weakness. FP-Growth uses the FP-Tree technique which is famous for the divide and conquer methods and does not generate itemset candidate generation. Fuzzy C-Covering uses the max item threshold technique to limit the execution of transactions. This algorithm requires a large memory and long execution time because of repeated data scans, since it is implemented to FP-Tree. The sales transaction data for IKK

more »

... e in 2018 amounted to 51,384 data. Data is used to identify items that might be purchased together with other items. Currently cooperatives do not have a data processing system for analysis of consumer buying patterns. Research is conducted to find association rules by implementing FP-Growth and Fuzzy C-Covering algorithms based on FP-Tree and to measure performance between algorithms based on execution time, memory usage, and the accuracy of association rules. Based on the test results, Fuzzy c-Covering based on FP-Tree uses less memory because the results of the tree formation are not stored and the execution time is longer because it is defined in the fuzzy set. FP-Growth has higher accuracy with the resulting association rules is risoles rahmat, tahu isi emly, pastel bihun susi with support 0.023%, and confidence 100%. Whereas Fuzzy c-Covering based on FP-Tree generates aqua 600ml, nasi telor balado siska, tahu bakso siska with support 0.05%, and confidence 21%.