Document Detail


Robustness metrics for measuring the influence of additive noise on the performance of statistical classifiers.
MedLine Citation:
PMID:  9315499     Owner:  NLM     Status:  MEDLINE    
Abstract/OtherAbstract:
This paper presents a novel quality measure called robustness. The robustness measure quantifies the influence of measurement noise in the attribute values on the credibility of the classification of a case. It is assumed that the type of distribution of the noise-generating process is known. It is not simple to measure the robustness in the general situation where the noise-free distribution of the attributes is unknown. Therefore, two approximations are suggested and compared with the robustness measure based on the noise-free distribution of the attributes. The usefulness of the suggested robustness measure is explored in a simulation experiment.
Authors:
M Egmont-Petersen; J L Talmon; A Hasman
Related Documents :
7373519 - Attention, unpredictability, and reports of physical symptoms: eliminating the benefits...
20968509 - Broadband active noise control with transfer path equalization.
17281609 - A method based on wavelet transform to reduce 1/f noise.
17952169 - Noise distribution of mueller matrices retrieved with active rotating polarimeters.
8829389 - Performances of surface laplacian estimators: a study of simulated and real scalp poten...
12786439 - Intrinsic noise-induced phase transitions: beyond the noise interpretation.
18087379 - Transmission reflection anomaly in second-harmonic generation from a monolayer.
7373519 - Attention, unpredictability, and reports of physical symptoms: eliminating the benefits...
6161799 - An automatic, optically isolated, biphasic constant current stimulator adapter for arte...
Publication Detail:
Type:  Journal Article    
Journal Detail:
Title:  International journal of medical informatics     Volume:  46     ISSN:  1386-5056     ISO Abbreviation:  Int J Med Inform     Publication Date:  1997 Sep 
Date Detail:
Created Date:  1997-10-28     Completed Date:  1997-10-28     Revised Date:  2004-09-03    
Medline Journal Info:
Nlm Unique ID:  9711057     Medline TA:  Int J Med Inform     Country:  IRELAND    
Other Details:
Languages:  eng     Pagination:  103-12     Citation Subset:  IM    
Affiliation:
Department of Biophysics, Maastricht University, The Netherlands. michael.egmont@bf.unimaas.nl
Export Citation:
APA/MLA Format     Download EndNote     Download BibTex
MeSH Terms
Descriptor/Qualifier:
Bayes Theorem*
Models, Statistical*
Quality Control
Reproducibility of Results

From MEDLINE®/PubMed®, a database of the U.S. National Library of Medicine


Previous Document:  TSMI: a CEN/TC251 standard for time specific problems in healthcare informatics and telematics.
Next Document:  Misunderstanding between knowledge engineers and domain experts: a re-evaluation of the use of induc...