What-if Analysis in MOLAP Environments
Guoliang Zhou, Hong Chen · 2009
What-if analysis is important for decision-maker to evaluate adopting new strategy and help knowledge discovery. It mostly occurs in OLAP environments, based on implementation OLAP is divided into MOLAP and ROLAP. Most research focus on ROLAP but in the industrial sector MOLAP is more common. There is relatively little work talking about what-if analysis in MOLAP environments. The challenge of what-if analysis is how to store hypothetical modified data and how these data efficiently participating cube calculation. In this paper we systemic analyze what-if analysis in MOLAP; hypothetical modified data stored as compression offset-data and is organized as HU-tree which is a variant of r*-tree. Through HU-tree we can efficient storage sparse hypothetical modified data which ordered by designed dimension order and fast retrieve. Then in cube calculation, hypothetical modified data and original data sort-merge join means to get result.