Abstracts - faqs.org

Abstracts

Engineering and manufacturing industries

Search abstracts:
Abstracts » Engineering and manufacturing industries

Why a best-fitting line might not be

Article Abstract:

A regular data analysis procedure such as fitting a staight line to bivariate (Y vs X) normally calls for the least squares solution. However, when situations like discrepant values in data sets come up, a new approach needs to be adopted. What is being proposed is to fit a line to most of the data regardless of the small percentage of values that do not adhere to the main pattern. This represents the point that majority of the data emanates from incontrol processes, but a small percentage also comes from out-of-control behaviour.

Author: Gunter, Bert
Publisher: American Society for Quality Control, Inc.
Publication Name: Quality Progress
Subject: Engineering and manufacturing industries
ISSN: 0033-524X
Year: 1992
Models, Correlation (Statistics), Least squares

User Contributions:

Comment about this article or add new information about this topic:

CAPTCHA


Tree-based procedures

Article Abstract:

The principles of tree-based classification are discussed. A classification tree is defined as a hierarchical set of questions on the taxonomy of an individual which allows the determination of the class which the individual belongs. Benefits of a classification tree include flexible construction, easy interpretation and application to continuous response data. Implementation of the technique in classifying plastic gears is described.

Author: Gunter, Bert
Publisher: American Society for Quality Control, Inc.
Publication Name: Quality Progress
Subject: Engineering and manufacturing industries
ISSN: 0033-524X
Year: 1998
Classification, Decision tree, Decision trees

User Contributions:

Comment about this article or add new information about this topic:

CAPTCHA


Fitting a three-group resistant line

Article Abstract:

An alternativeto least squares in fitting a line to X-Y data is the three-group median procedure by statistician John Tukey. Tukey's method is slightly modified to fit a line to 17 X-Y pairs with an overall linear trend. Repeated iteration is applied to achieve negligible changes and residuals are obtained from intercepts until the fit is adequate.

Author: Gunter, Bert
Publisher: American Society for Quality Control, Inc.
Publication Name: Quality Progress
Subject: Engineering and manufacturing industries
ISSN: 0033-524X
Year: 1993
Iterative methods (Mathematics), Usage, Iteration (Mathematics), Column, Curve fitting

User Contributions:

Comment about this article or add new information about this topic:

CAPTCHA


Subjects list: Methods
Similar abstracts:
  • Abstracts: Another meeting? At the right time and for the right reason? It's really all about you
  • Abstracts: Making light work. With filters or without. Designer demands
  • Abstracts: Why microfilm is still a winner. Making a virtue of the unthinkable. Look, listen and learn
  • Abstracts: Designing for zero leakage. 'Thinking' bearings just round the corner. Reducing friction is just the start
  • Abstracts: 'Live' tooling breaks design barriers. Real returns in design solutions
This website is not affiliated with document authors or copyright owners. This page is provided for informational purposes only. Unintentional errors are possible.
Some parts © 2025 Advameg, Inc.