Association rule mining method, system and device under low resource conditions

Abstract

The invention belongs to the field of data mining, and specifically relates to an association rule mining method, system, and device under low-resource conditions, aiming to solve the problem that existing association rule mining methods cannot complete low-frequency association rule mining under low memory conditions. The method of the system includes: obtaining the data set of association rules to be mined, digitally encoding the items in each transaction of the data set; obtaining high-frequency frequent itemsets and high-frequency association rules; constructing the first low-frequency frequent itemset and the second low-frequency frequent Itemset: Based on the first low-frequency frequent itemset and the second low-frequency frequent itemset, the third low-frequency frequent itemset is obtained through the setdiff function, and the low-frequency association rules are obtained according to the third low-frequency frequent itemset; the high-frequency association rules and low-frequency association are output rule. The invention can mine complete low-frequency association rules under low-memory conditions.

Description

technical field [0001] The invention belongs to the field of data mining, and in particular relates to an association rule mining method, system and device under low resource conditions. Background technique [0002] In some cases, such as personal health data, we need to pay attention to the relationship between the data of interest, such as how some symptoms and diseases are related. In order to meet this requirement, we mine association rules on the data set. [0003] Association rule mining is an important data mining algorithm that can discover interesting relationships in a database. Its purpose is to use some metrics to distinguish the strong rules that exist in the database. Two commonly used association rule algorithms are A-priori algorithm and FPGrowth algorithm. The A-priori algorithm uses the method of generating candidate pairs to scan the data set multiple times, and each generated candidate pair requires a large amount of memory, especially when the suppor...

AI Technical Summary

Problems solved by technology

[0007] In order to solve the above-mentioned problems in the prior art, that is, in order to solve the problem that the existing association rule mining methods cannot complete the mining of low-frequency association rules under low memory conditions, the first aspect of the present invention proposes a low-resource Association rule mining method, the method includes:

Images