AbstractItemset mining is a popular extension to the frequent pattern mining problem in data mining. Finding infrequent patterns, however, has gained its importance due to proven utility in the areas of web mining, bioinformatics and others. High utility mining refines the problem focus to identifying business-relevant transaction patterns that take purchase quantities and monetary considerations into account, like unit price and cost, typically to identify patterns of profit potential. High utility infrequent itemset mining unveils rare cases of highly profitable itemsets. This paper proposes a customized Ant colony algorithm for the efficient discovery of high utility infrequent itemsets. The mining performance of proposed algorithm is analyzed on four real time datasets namely chess, food mart, mushroom and retail.

中文翻译：

---

使用定制的蚁群算法进行高效用的不频繁项集挖掘

AbstractItemset 挖掘是数据挖掘中频繁模式挖掘问题的流行扩展。然而，由于在网络挖掘、生物信息学和其他领域中已被证明的实用性，发现不常见的模式变得越来越重要。高效用挖掘将问题重点细化到识别业务相关的交易模式，这些交易模式考虑了购买数量和货币因素，如单价和成本，通常用于识别潜在的利润模式。高效用的非频繁项集挖掘揭示了高利润项集的罕见情况。本文提出了一种定制的蚁群算法，用于高效发现高效用的不频繁项集。在国际象棋、食品市场、蘑菇和零售四个实时数据集上分析了所提出算法的挖掘性能。