Intelligent Analysis of Data Cube via Statistical Methods

Data cube is a multi-dimensional structure of data representation used in data warehousing, which is analyzed using Online Analytical Process( OLAP). However, OLAP in itself is incapable of intelligent analysis in terms of generating a compact and useful data cube as well as lacks the power of predicting empty measures in data cube. Recently statistical methods have been applied for compact cube generation and prediction of empty measures in data cube. In this paper we reviewed and critically evaluated the available work done in this area. Literature review highlighted that sparsity, data redundancy and need of domain knowledge are problems in the way of intelligent analysis. Statistical methods have solved these issues in schema generation, compact cube generation and in prediction of empty measures of data cube individually. Methodology to solve these problems and then to perform an intelligent analysis has not yet been developed. In order to develop such an integrated methodology, we propose a conceptual model for intelligent analysis of data cube via statistical methods in the first stage. This model is integrating statistical methods used in generating a compact cube with a prediction mechanism. Our next target is to develop this model into a complete methodology.