NotesFAQContact Us
Search Tips
ERIC Number: ED356259
Record Type: Non-Journal
Publication Date: 1992
Pages: 10
Abstractor: N/A
Reference Count: N/A
Least Principal Components Analysis (LPCA): An Alternative to Regression Analysis.
Olson, Jeffery E.
Often, all of the variables in a model are latent, random, or subject to measurement error, or there is not an obvious dependent variable. When any of these conditions exist, an appropriate method for estimating the linear relationships among the variables is Least Principal Components Analysis. Least Principal Components are robust, consistent, and sufficient maximum likelihood estimates of the best total linear fit to observed data. They are more appropriate than regression estimates when the smallest eigenvalue exists and is distinct from the next smallest, and the variability to minimize is in more than one variable or when multicollinearity is a problem. They are as easy to compute as are common principal components because they are principal components. T. W. Anderson (1963) provides a theory of inferential statistics for Principal Components that can be used in computing significance levels and confidence intervals for least principal components as well. Bootstrap approaches have also been developed. (SLD)
Publication Type: Reports - Evaluative; Speeches/Meeting Papers
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A