Scholarlink

<p>The reverse Skyline query aims to identify a set of data points that dynamically dominate a query point from the decision-makers’ perspective. It has been widely applied in business decision-making, recommendation systems, location-based services, knowledge discovery, and data mining. However, existing reverse Skyline queries mainly focus on single-point queries, overlooking multi-point combination queries. To address this, we propose the concept of combination reverse Skyline query (CRSQ), based on single query combinations. Furthermore, to handle multiple combinations with different cardinalities, we develop the Multiple Enhanced <i>k</i> combination reverse Skyline query method (M<i>k</i>ECRSQ). M<i>k</i>ECRSQ includes three main phases. Initially, we prove that <i>k</i> combination reverse Skyline query (<i>k</i>CRSQ) is NP-hard and propose a novel index structure called QR-GMap for combination queries to significantly accelerate <i>k</i>CRSQ. Subsequently, we compare the multiple <i>k</i>CRSQ results of various <i>k</i> values to determine the most dominant combinations. Finally, we expand the result set by proving the monotonicity of the ECRSQ algorithm. The final M<i>k</i>ECRSQ results consist of the obtained combinations and the expanded result set. Theoretical and experimental results show that M<i>k</i>ECRSQ not only rapidly yields results for CRSQ but also recommends the most dominant combinations to decision-makers among multiple combinations in the query dataset, while also overcoming the challenge of limited cardinality in the result sets. By introducing CRSQ and M<i>k</i>ECRSQ, our work fills a significant research gap in reverse Skyline queries, extending their applicability to multi-point combination queries and offering enhanced decision-making support.</p>

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

Research on multiple enhanced k combination reverse Skyline query method

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年1月28日

查看全文
  • Song Li,
  • Xinyuan Zhang,
  • Liping Zhang

摘要

The reverse Skyline query aims to identify a set of data points that dynamically dominate a query point from the decision-makers’ perspective. It has been widely applied in business decision-making, recommendation systems, location-based services, knowledge discovery, and data mining. However, existing reverse Skyline queries mainly focus on single-point queries, overlooking multi-point combination queries. To address this, we propose the concept of combination reverse Skyline query (CRSQ), based on single query combinations. Furthermore, to handle multiple combinations with different cardinalities, we develop the Multiple Enhanced k combination reverse Skyline query method (MkECRSQ). MkECRSQ includes three main phases. Initially, we prove that k combination reverse Skyline query (kCRSQ) is NP-hard and propose a novel index structure called QR-GMap for combination queries to significantly accelerate kCRSQ. Subsequently, we compare the multiple kCRSQ results of various k values to determine the most dominant combinations. Finally, we expand the result set by proving the monotonicity of the ECRSQ algorithm. The final MkECRSQ results consist of the obtained combinations and the expanded result set. Theoretical and experimental results show that MkECRSQ not only rapidly yields results for CRSQ but also recommends the most dominant combinations to decision-makers among multiple combinations in the query dataset, while also overcoming the challenge of limited cardinality in the result sets. By introducing CRSQ and MkECRSQ, our work fills a significant research gap in reverse Skyline queries, extending their applicability to multi-point combination queries and offering enhanced decision-making support.