Outlier Detection Through Bipartite Visual Analytics

Details

• Personal Author:
  Bhavnani SK ; Dang B ; Drake JA ; Vishweswaran S
• Description:
  A critical goal of outlier detection is to determine whether an outlying value was caused by experimental/human error, or by natural biological diversity. However, because univariate or multivariate methods (e.g., box plots and principle component analysis) typically used for outlier detection use unipartite representations, they cannot distinguish whether outliers across a set of variables represent, for example, a single patient or different patients. Here we propose a bipartite visual analytical approach to outlier detection, and demonstrate its usefulness for identifying complex bipartite outliers in a dataset of rickettsioses patients, which enabled domain experts to determine whether the outliers were caused by errors, or by biological diversity. [Description provided by NIOSH]
• Subjects:
  Decision Making Occupational Health Safety
• Keywords:
  Biological Distribution Biological Factors Decision Making Statistical Analysis Statistical Quality Control
• ISSN:
  2153-4063
• Document Type:
  Text
• Funding:
  Grant
• Genre:
  Abstract or Summary
• Place as Subject:
  OSHA Region 3 ; OSHA Region 6 ; Pennsylvania ; Texas
• CIO:
  National Institute for Occupational Safety and Health (NIOSH)
• Topic:
  Workplace Safety & Health
• Location:
  North America
• Pages in Document:
  20
• Volume:
  2013
• NIOSHTIC Number:
  nn:20055947
• Citation:
  AMIA Jt Summits Transl Sci Proc 2013 Mar; 2013:20
• Federal Fiscal Year:
  2013
• NORA Priority Area:
  Services
• Performing Organization:
  University of Texas Medical Branch, Galveston
• Peer Reviewed:
  False
• Start Date:
  20100901
• Source Full Name:
  AMIA Joint Summits on Translational Science Proceedings
• End Date:
  20130131
• Collection(s):
  National Institute for Occupational Safety and Health
• Main Document Checksum:
  urn:sha-512:d061a2537b6c4d670982edb221d1b9b8ef0fa70ab0bf07670968697d36086691270e43aa374d4d59a6c3b1dfb8e90037587bdeb28903273ff3757f2982180991
• Download URL:
  https://stacks.cdc.gov/view/cdc/213694/cdc_213694_DS1.pdf
• File Type:
  [PDF - 405.58 KB ]