Explanation of Exceptional Values in Multi-dimensional Business Databases Defended on Thursday, 14 November 2013
Multi-dimensional or OnLine Analytical Processing (OLAP) databases are a popular business intelligence (BI) technique in the field of enterprise information systems for business analytics and decision support. In this dissertation, OLAP database functionality is extended to support the business analyst in the exploration of OLAP data. The database is augmented with novel functionality for the detection of exceptional values, explanation generation, and sensitivity analysis. We describe how exceptional values at any level in the data, can be automatically detected by statistical and managerial models. It is also shown how exceptional values can be explained by underlying causes. This is realized by a generic model for diagnosis of atypical values. By applying it, a full explanation tree of causes at successive levels can be generated. If the tree is too large, the analyst can use appropriate filtering measures to prune the tree to a manageable size. The purpose of the methods and algorithms presented here, is to provide OLAP databases with more powerful explanatory analytics and reporting functions. This methodology has a wide range of applications, such as variance analysis in accounting, competition benchmarking, analysis of sales and financial data, and the analysis of any other data that possess a multi-dimensional hierarchical structure. The method is demonstrated in several case studies. For example, the explanatory analysis of a sales data cube is discussed, and computerized competition benchmarking with financial data about Dutch retail companies is illustrated.
business intelligence; business analytics; multi-dimensional databases; decision-support systems; OLAP; exception reporting; explanation; variance analysis; sales analysis; finance & accountancy