Chapter 15
Orthogonal Polynomial and Treatment Quantification for Missing Data


Data goes missing when test samples fall, experimental data is recorded incorrectly, measurements are incorrect, samples don't match factor level settings, or when there is no measurable quantity for the output responses, etc. In some cases, the test samples are made appropriately but the testing equipment does not function correctly (for example, there is a driving motor malfunction or a copy machine does not work as intended). In some reliability or durability tests, the output response data is not obtained because the samples don't yield under the loads within the test duration.

If samples are correctly made and the data is missing because of insufficient functionality of the target system, it is common to apply reasonable replacement values for the missing data. For example, you can use the minimum measured S/N (signal-to-noise) value minus 3 (dB) to estimate the missing S/N ratio data due to insufficient functionality [or the maximum measured S/N value plus 3 (dB) for excessive functionality such as surviving a durability].

This chapter illustrates how to calculate reasonable replacement values as estimates for data that's missing because of any number of experimental incidents.

  • 15.1 Missing Data
  • 15.2 Replacement Values for Missing Data
  • 15.3 Comparison Between Approximation Methods Based on Main-Effect Plots and Orthogonal Polynomials
  • 15.4 Iterative Approximations When There are Multiple Missing Values
  • 15.5 Quantification of Missing Data Caused by Various Test Disruptions

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In