A Design Science Framework for Developing Data Mining Artifacts

Abstract:

Extracting  valuable  information  from  digital  data  is  known  as  Knowledge  Discover  in Databases (KDD) and objective of Information Systems (IS) research. Data Mining (DM), the extraction and verification of hidden patterns in data, is the core activity of the KDD process. Literature relating to DM is strongly focused on scientific rigor of the research, however, by contrast part of the IS community finds practical relevance lacking in DM research. Based on a literature review of DM research paradigms, this article develops a data mining approach based  on  design  science theory.  The  application  of  this  approach  on  a  real-case  business problem  reveals  the  acceptance  of  different forms  of  data  and  the  consideration  of  the interactive and iterative nature of DM key success factor for DM projects. Moreover, business needs are confining the necessary depth of DM activities. Building on the evaluation of the model within the case study, we suggest the results might point out an explicit connection between  business  and  science especially  in  the  field  of  DM.  This  paper  provides  a contribution to both the DM professionals and the DM community in IS research. Further, our approach should be examined and verified for other DM related practical cases as well, as it might contribute to a unifying theory of DM based on design science research.