Scholarlink

In this paper, we introduce ClassCube, a novel methodology designed to perform efficient and effective classification over large, multidimensional OLAP data cubes. ClassCube leverages logical cuboid lattices to represent data across multiple aggregation levels, enabling intelligent selection of both dimensions and cuboids for classification tasks. By integrating dimensionality reduction techniques such as Dimension Selection and Principal Component Analysis, the approach significantly reduces computational overhead while maintaining high classification accuracy. Extensive experimental assessments confirm that ClassCube achieves an optimal balance between efficiency and accuracy, highlighting its suitability for real-life big data analytics applications.

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

ClassCube: Effective and Efficient Big OLAP Data Cube Classification via Dimensionality Reduction

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Alfredo Cuzzocrea,
  • Mojtaba Hajian,
  • Abderraouf Hafsaoui

摘要

In this paper, we introduce ClassCube, a novel methodology designed to perform efficient and effective classification over large, multidimensional OLAP data cubes. ClassCube leverages logical cuboid lattices to represent data across multiple aggregation levels, enabling intelligent selection of both dimensions and cuboids for classification tasks. By integrating dimensionality reduction techniques such as Dimension Selection and Principal Component Analysis, the approach significantly reduces computational overhead while maintaining high classification accuracy. Extensive experimental assessments confirm that ClassCube achieves an optimal balance between efficiency and accuracy, highlighting its suitability for real-life big data analytics applications.