Efficient cube computing on an extended multidimensional model over uncertain data

Data uncertainty is an inherent property in various applications due to reasons such as measurement errors, incompleteness of data and so on. While On-Line Analytical Processing (OLAP) has been a powerful method for analyzing large data warehouse, OLAP over uncertain data has become a valuable and attractive issue because of the increasingly demand for handling uncertainty in multidimensional data. In this paper, we firstly describe our UStar-Schema model that extends the traditional OLAP model to support uncertain dimension attributes in fact table, uncertain measures in fact table and uncertainty in dimension table. Then we extend the processing model of the aggregate queries and cube computing on Ustar-Schema. Secondly, we design a novel index structure called PSI-Index on UStar-Schema to improve efficiency of OLAP quering and cube computing. Furthermore, an advanced index structure called HB-Index and an efficient algorithm are proposed to accelerate iceberg cube computing based on our model using pruning techniques to eliminate huge amounts of useless computations. Finally, extensive experiments are performed to examine the efficiency and effectiveness of our proposed techniques.