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

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.