The item set based on the node list indicates the latest progress of the framework's frequent item set mining and the latest progress of the framework.
The latest paper was published in Expert Systems with Applications 2015, volume 42, Issue 13.
This paper uses the equivalence class improvement strategy, which greatly improves the mining speed and saves memory consumption. The PrePost + algorithm is superior to the PrePost and FIN algorithms in terms of time and space performance.
PrePost + algorithm: http://www.cis.pku.edu.cn/faculty/system/dengzhihong/Source%20Code/prepost+.cpp
Related papers:
Http://www.sciencedirect.com/science/article/pii/S0957417415001803
Because the expression of the node list greatly reduces the time required to support the computing item set from the underlying layer, PrePost +, PrePost, and FIN can be expanded to mine Rare Itemsets, Closed Itemsets, and Maximal Itemsets, frequent Generator Itemsets, Frequent Itemsets from Data Stream, Frequent Itemsets from Uncertain Data, etc.
In addition, these three algorithms are particularly suitable for modifying to parallel frequent mode mining algorithms (such as GPU-based, Hadoop-based, and other parallel computing platforms ). Interested peers can conduct research on the above work.