Process prediction in noisy data sets: a case study in a Dutch hospital

Research output: Chapter in Book/Report/Conference proceedingChapterAcademicpeer-review

22 Citations (Scopus)
7 Downloads (Pure)

Abstract

Predicting the amount of money that can be claimed is critical to the effective running of an Hospital. In this paper we describe a case study of a Dutch Hospital where we use process mining to predict the cash flow of the Hospital. In order to predict the cost of a treatment, we use different data mining techniques to predict the sequence of treatments administered, the duration and the final ‿care product‿ or diagnosis of the patient. While performing the data analysis we encountered three specific kinds of noise that we call sequence noise, human noise and duration noise. Studies in the past have discussed ways to reduce the noise in process data. However, it is not very clear what effect the noise has to different kinds of process analysis. In this paper we describe the combined effect of sequence noise, human noise and duration noise on the analysis of process data, by comparing the performance of several mining techniques on the data.
Original languageEnglish
Title of host publicationData-Driven Process Discovery and Analysis
Subtitle of host publicationSecond IFIP WG 2.6, 2.12 International Symposium, SIMPDA 2012, Campione d’Italia, Italy, June 18-20, 2012, Revised Selected Papers
EditorsPhilippe Cudre-Mauroux, Paolo Ceravolo, Dragan Gašević
Place of PublicationBerlin
PublisherSpringer
Pages60-83
Number of pages24
ISBN (Electronic)978-3-642-40919-6
ISBN (Print)978-364240918-9
DOIs
Publication statusPublished - 2013
Event2nd IFIP WG 2.6, 2.12 International Symposium on Data-Driven Process Discovery and Analysis, SIMPDA 2012 - Campione d’Italia, Italy
Duration: 18 Jun 201220 Jun 2012
Conference number: 2

Publication series

NameLecture Notes in Business Information Processing
PublisherSpringer
Volume162
ISSN (Print)1865-1348
ISSN (Electronic)1865-1356

Conference

Conference2nd IFIP WG 2.6, 2.12 International Symposium on Data-Driven Process Discovery and Analysis, SIMPDA 2012
Abbreviated titleSIMPDA
CountryItaly
CityCampione d’Italia
Period18/06/1220/06/12

Keywords

  • IR-87303
  • METIS-297828
  • EWI-23700

Fingerprint Dive into the research topics of 'Process prediction in noisy data sets: a case study in a Dutch hospital'. Together they form a unique fingerprint.

Cite this