Resumen
Association rules mining progresses year by year. There are many algorithms of association rules mining. The most popular are the Apriori algorithm and the FP-Growth algorithm. But these algorithms have disadvantages. The Apriori algorithm requires many transaction base passes. The FP-Growth algorithm uses a many-edged (non-binary) tree data structure. Algorithm is characterized by the data structure used in it. We discover an association rules mining algorithm using three-dimensional data structure. Algorithm needs only two transaction base passes. The first pass is to insert transactions in three-dimensional data structure. The second pass is to count support of extracted from three-dimensional data structure itemsets. The algorithm tested and compared with the Apriori algorithm and the FP-Growth algorithm. The algorithm is more effective by memory usage than the FP-Growth algorithm, when number of unique elements is between 10 and 868, and the Apriori algorithm, when number of unique elements is between 49 and 498.