A framework for data mining and KDD

  • Authors:
  • Ingolf Geist

  • Affiliations:
  • University of Magdeburg, D-39016 Magdeburg, Germany

  • Venue:
  • Proceedings of the 2002 ACM symposium on Applied computing
  • Year:
  • 2002

Quantified Score

Hi-index 0.00

Visualization

Abstract

The KDD process is a non-trivial, iterative, interactive and multi-step process, that requires the development of a unifying model. This model have to ensure an uniform description of data and patterns and the control of the manipulation of the data and patterns. Thus, the model defines operations within the pattern and data, as well as transition operations between data and patterns.This paper proposes a framework consisting of a model view, a data view and a process view. It focuses on the model and data view. The model view contains a set of mining models, which contain all information of a data mining result, that are based on constraints. The proposed model algebra uses concepts of constraint databases as well as collective and parallel data mining. The whole process is supported by using operations between data and model view.