Scholarlink

The need for efficient pattern extraction techniques has increased due to the rapid expansion of data in a variety of industries, including healthcare, e-commerce, and entertainment. Because of their low support values, unusual but important itemsets are frequently missed by traditional association rule mining (ARM) algorithms. This work suggests a novel approach that uses item-specific minimum support (MIS) restrictions to generate precise uncommon association rules with 100% confidence. In contrast to traditional methods, the suggested method effectively mines uncommon patterns without generating an excessive amount of candidate itemsets by utilizing both Predefined Item Support (PIS) and MIS values. Experiments on real-world datasets, such as movie ratings and medical records, show that the method performs better than the Apriori algorithm in terms of rule generation efficiency and execution time. The findings demonstrate its potential for revealing high-impact, hidden correlations that conventional frequent pattern mining tools frequently overlook.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

An Exact Rare Rule Mining Approach Using Minimum Item Support Constraint from Heterogeneous Data

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Sudarsan Biswas,
  • Diganta Saha,
  • Rajat Pandit,
  • Neepa Biswas

摘要

The need for efficient pattern extraction techniques has increased due to the rapid expansion of data in a variety of industries, including healthcare, e-commerce, and entertainment. Because of their low support values, unusual but important itemsets are frequently missed by traditional association rule mining (ARM) algorithms. This work suggests a novel approach that uses item-specific minimum support (MIS) restrictions to generate precise uncommon association rules with 100% confidence. In contrast to traditional methods, the suggested method effectively mines uncommon patterns without generating an excessive amount of candidate itemsets by utilizing both Predefined Item Support (PIS) and MIS values. Experiments on real-world datasets, such as movie ratings and medical records, show that the method performs better than the Apriori algorithm in terms of rule generation efficiency and execution time. The findings demonstrate its potential for revealing high-impact, hidden correlations that conventional frequent pattern mining tools frequently overlook.