Survey on techniques of High Utility Mining

computer Engg.

Frequent Itemset mining is essential in data mining and its popular application is market basket analysis. Traditional FIM model produce large number of frequent but low revenue itemset and lose information on valuable itemset with low selling frequencies. Utility mining is essential to avoid this. Discovery of high utility itemset (e.g. high profit,weight) from database is an important task. However, discovered HUIs may not be compact in size which degrades efficiency of mining process. To achieve efficiency, some compact representations available such as free sets, non-derivable sets and closed itemsets. Mining closed high utility itemsets serves as a compact and lossless representation of HUIs. To achieve this, algorithm CHUD (Closed High Utility Itemset Discovery) is proposed in literature. Organizations mine huge data to generate HUI. Organizations conducting business have to suffer the huge data mining computation and storage overhead. To avoid the computation burden of mining enormous data, there is a scope to outsource this task to another party. But outsourced party may not be trustworthy so there’s the need of preservation of data privacy from outsourced party. Various privacy preservation techniques are available in literature. And instead of mining whole database again for getting HUIs, by using techniques from [10] we can also recover all HUI’s from the concise representation. In this paper we have studied different techniques available for the compact representation of HUIs.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.