Extending QMBE Language with Clustering

Extending QMBE Language with Clustering

Ana Azevedo (Polythecnic Institute of Porto/ISCAP, Portugal & Algoritmi R&D Center, University of Minho, Guimarães, Portugal) and Manuel Filipe Santos (Algoritmi Center, Department of Information Systems, University of Minho, Guimarães, Portugal)
Copyright: © 2013 |Pages: 19
DOI: 10.4018/ijdsst.2013100104


Business Intelligence (BI) is an important area of the Decision Support Systems (DSS) discipline. Over the past years, the evolution in this area has been considerable. Similarly, in the last years, there has been a huge growth and consolidation of the Data Mining (DM) field. DM is being used with success in BI systems, but a truly DM integration with BI is lacking. This creates a gap between DM and BI systems. With the purpose of closing this gap a new DM language for BI, named as Query-Models-By-Example (QMBE), was envisaged and implemented with success, but addressing only classification rules. This paper presents an extension of QMBE language to include clustering. This represents one more step towards the integration of DM with BI, which constitutes an important issue.
Article Preview


Business Intelligence (BI) can be presented as an architecture, tool, technology or system that gathers and stores data, analyzes it using analytical tools, and delivers information and/or knowledge, facilitating reporting, querying, and, ultimately, allows organizations to improve decision making (Clark, Jones, & Armstrong, 2007; Kudyba & Hoptroff, 2001; Michalewicz, Schmidt, Michalewicz, & Chiriac, 2007; Moss & Shaku, 2003; Negash, 2004; Raisinghani, 2004; Thierauf, 2001; Turban, Sharda, & Delen, 2011). To put it shortly, Business Intelligence can be defined as the process that transforms data into information and then into knowledge (Golfarelli, Rizzi, & Cella, 2004). Being rooted in the Decision Support Systems (DSS) discipline, BI has suffered a considerable evolution over the last years and is, nowadays, an area of DSS that attracts a great deal of interest from both the industry and researchers (Arnott & Pervan, 2008; Clark, Jones, & Armstrong, 2007; Hannula & Pirttimäki, 2003; Hoffman, 2009; Negash, 2004; Richardson, Schlegel, & Hostmann, 2009; Richardson, Schlegel, Hostmann, & McMurchy, 2008).

Data Mining (DM) is being applied with success in BI and several examples of applications can be found (Linoff, 2008; Turban, Sharda, & Delen, 2011; Vercellis, 2009). Despite that, DM has not yet reached to non-specialized users. The authors consider that the real issue is related with the fact that the Knowledge Discovery in Databases (KDD) process, as presented by Fayyad, Piatetski-Shapiro, and Smyth (1996), is not fully integrated in BI. Consequently, its full potential could be not completely explored by decision makers using the systems. Currently, DM systems are functioning as separate isles.

The authors have point out three main reasons for DM to be not completely integrated with BI. Firstly, the models/patterns obtained from DM are complex and there is the need of an analysis from a DM specialist. This fact can lead to a non-effective adoption of DM in BI, being that DM is not really integrated on most of the implemented BI systems, nowadays. Secondly, the problem with DM is that there is not a user-friendly tool that can be used by decision makers to analyze DM models. Usually, BI systems have user-friendly analytical tools that help decision makers in order to obtain insights on the available data and allow them to take better decisions. Examples of such tools are On-Line Analytical Processing (OLAP) tools, which are widely used (Negash, 2004; Turban, Sharda, & Delen, 2011). Powerful analytical tools, such as DM, remain too complex and sophisticated for the average consumer. Finally, but extremely important, it has not been given sufficient emphasis to the development of solutions that allow the specification of DM problems through business oriented languages, and that are also oriented for BI activities. With the expansion that has occurred in the application of DM solutions in BI, this is, currently, of increasing importance.

Complete Article List

Search this Journal:
Open Access Articles
Volume 11: 4 Issues (2019): 1 Released, 3 Forthcoming
Volume 10: 4 Issues (2018)
Volume 9: 4 Issues (2017)
Volume 8: 4 Issues (2016)
Volume 7: 4 Issues (2015)
Volume 6: 4 Issues (2014)
Volume 5: 4 Issues (2013)
Volume 4: 4 Issues (2012)
Volume 3: 4 Issues (2011)
Volume 2: 4 Issues (2010)
Volume 1: 4 Issues (2009)
View Complete Journal Contents Listing