Association Rule Mining Algorithm of Transposed matrix Based on Python Language

摘要

Apriori and its improved algorithms can be generally classified into two kinds: SQL-based and on memory-based. In order to improve association rule mining efficiency, after analyzing the efficiency bottlenecks in some algorithms of the second class, an improved efficient algorithm for Python language is proposed. Two matrixes are introduced into the algorithm: one is used to map database and the other to store frequent 2-itemsets related information. Through the operation of two matrixes, its time complexity and space complexity decrease significantly. The experiment indicates that the method has better performance.