Journal of Accounting and Management Information Systems (JAMIS)

Business Intelligence – types of applications

19/2007 ,   p59..72

Robert ŞOVA
Anamaria ŞOVA

Keywords:   Business Intelligence, KDD, Data mining, Reporting, Ad-hoc query

In recent years, new Business Intelligence (BI) technologies are emerging quickly. The main objective of the BI technologies is to help people understand data more quickly with the purpose of supporting better and faster decision-making and, finally, attain business objectives. The BI technologies may increase organizational efficiency and effectiveness. BI technologies are not the end game. Falling into that mind-set can get you into real trouble - trouble with ROI. The end game is the business process efficiencies and effectiveness that come from being able to make better decisions, faster. For this reason a strategic alignment of the BI technology with organization’s objectives is essential. But this is not enough. A good knowledge of the characteristics of each type of business intelligence applications is necessary for successfully using state-of-the-art BI technologies. Today’s enterprise organizations require a variety of different BI applications serving many different user communities. These applications generally fit into one of five categories, reporting applications, ad-hoc query and reporting, multidimensional analysis, model based applications and KDD applications. Reporting applications tend to provide static or parameterized reports. The audience tends to be broad. Reporting applications that have minimal analytical requirements are typically based on relational databases and use SQL as query language. Ad-hoc query and reporting offer the user a high level of interaction by providing a variety of data selection and navigation techniques. These applications are typically based on relational databases but often they are based on a multidimensional data model which offer a limited, but very useful, set of analytic capabilities. Multidimensional analysis applications also support ad-hoc exploration of data; however, they answer much more complex questions. This query is multidimensional, that is, calculations in the query occur along more than one dimension. For this, a multidimensional data files (the cube) must be created. Model based applications allow the user to predict outcomes or results. These applications generate new data using analytical tools such as models, forecasts, allocation methods and scenario management tools. KDD applications refer to the broad process of finding knowledge in data in the context of large databases. KDD process is interactive and iterative and comprises a several basic steeps. KDD refers to the over-all process of discovering useful knowledge from data while data mining refers to the applications of algorithms for extracting patterns from data. The additional steps of KDD process are essential to ensure that useful information (knowledge) is derived from data. Data mining involves fitting models to, or determining patterns from observed data. This paper offers o brief overview of data mining methods. Most data mining methods are based on concepts specific to machine learning, pattern recognition and statistics: classification, clustering, regression and so forth. The criteria for selecting applications can be divided into practical and technical. Practical criteria include consideration of the potential for significant impact of an application, no good alternatives exist, organizational support. Technical criteria include considerations such as the availability of sufficient data, relevance of attributes and the most important is the prior knowledge.