Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
IDENTIFICATION AND REMOVAL OF OUTLIERS IN DATA SETS
Document Type and Number:
WIPO Patent Application WO/2015/117907
Kind Code:
A3
Abstract:
The invention relates to a selectively-adaptive data processor which is designed to receive a plurality of sets of values, each set being assigned to a particular sample and containing both a data value of the data series to be processed and additional values that allow more precise assessment of this data value or of the data series. The data processor comprises a filter processor designed to selectively process a data series based on the sets received and issue this as a smoothed version, with a data value of an implausible set being replaced by an internally determined data value prior to data smoothing. Implausible sets are considered those sets for which the data value deviates strongly from the preceding curve of the data series or for which the data value is classified as being a data value of too low quality based on the additional values of the set.

Inventors:
KRENZER DANIEL (DE)
HESS ALBRECHT (DE)
KÁTAI ANDRÁS (DE)
Application Number:
PCT/EP2015/052011
Publication Date:
October 01, 2015
Filing Date:
January 30, 2015
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
FRAUNHOFER GES FORSCHUNG (DE)
International Classes:
G06K9/00; G06F17/18; G06T7/20
Foreign References:
US20130346844A12013-12-26
US20070066916A12007-03-22
Other References:
JOHN LUDBROOK: "OUTLYING OBSERVATIONS AND MISSING VALUES: HOW SHOULD THEY BE HANDLED?", CLINICAL AND EXPERIMENTAL PHARMACOLOGY AND PHYSIOLOGY, vol. 35, no. 5-6, 1 May 2008 (2008-05-01), pages 670 - 678, XP055190361, ISSN: 0305-1870, DOI: 10.1111/j.1440-1681.2007.04860.x
Attorney, Agent or Firm:
HERSINA, Günter et al. (Zimmermann Stöckeler, Zinkler, Schenk & Partner mb, Radlkoferstrasse 2 München, DE)
Download PDF: