Affordable Access

Publisher Website

Detecting outliers in multivariate process data by using convex hulls

DOI: 10.1016/s1570-7946(00)80019-x
  • Computer Science


In this paper the problem of near-real-time detection and removal of up to 15% radial outliers from large data sets (10000 or more records) is investigated and a practical solution demonstrated. The proposed method was compared to the Rocke and Woodruff algorithm in an example of elliptically arranged random data and applied successfully to test data recorded of a flotation plant. The technique has the benefits of low computational cost with minimal operator input, and can be implemented as a real-time outlier detection tool.

There are no comments yet on this publication. Be the first to share your thoughts.