Entwicklung und Anwendung eines Werkzeugs zur Klassifikation von Daten nach dem KDD-Prozess

Translated title of the contribution: Development and Application of a Tool for Classifying Data according to the KDD Process

Rene Vogl

Research output: ThesisMaster's / Diploma thesis

Abstract

Classication is a widespread and often-used data mining task. The steps involved in performing a data mining task are summarized and described in the KDD-process. Detailed descriptions of these steps and explanations of the dependencies among them are given in this thesis. The task of this thesis is to divide projects into categories according to the success of these projects. Based on this classication, it is possible to estimate the expected success of ongoing projects. The goal is to identify ongoing projects, which show evidence that they can not be completed successfully. Performing the actual classification of projects should be as easy and quick as possible. The result of this thesis is a tool that facilitates applying the KDD-process in order to solve classication tasks, such as the task to divide projects into categories according to the expected success. This tool allows to quickly and easily perform classication tasks without requiring users to have in-depth knowledge about the KDD process. The classification algorithms and techniques underlying the tool were chosen such that high-quality classification results are achieved even if only little data is available. The developed tool has been used to solve the given task of classifying projects according to their expected success and the quality of the results produced by the supported classification algorithms have been evaluated according to several criteria.
Translated title of the contributionDevelopment and Application of a Tool for Classifying Data according to the KDD Process
Original languageGerman (Austria)
Supervisors/Reviewers
  • Schrefl, Michael, Supervisor
  • Karlinger, Michael, Co-supervisor
Place of PublicationLinz
Publication statusPublished - Oct 2013

Fields of science

  • 102 Computer Sciences
  • 102015 Information systems
  • 502 Economics
  • 509 Other Social Sciences

JKU Focus areas

  • Computation in Informatics and Mathematics
  • Management and Innovation

Cite this