“A comparison of two methods for estimating parameters restricted by linear inequalities in linear statistical models was made.
The first method consisted of finding the least squares estimates of the parameters disregarding the restrictions, and then forcing those parameters that were outside of their allowable ranges to the nearest endpoints of those ranges.
The second method made use of quadratic programming to minimize the same sum of squares that the first method minimized while keeping the parameters inside or at the endpoints of their allowable ranges.
The quadratic programming gave better estimates of the parameters in the sense that the mean square error between the estimates and the true values of the parameters was smaller, but that method took much more computer time and more computer storage space"--Abstract, page ii.
Antle, Charles E.
Johnson, Charles A.
Anderson, Richard A.
Fuller, Harold Q., 1907-1996
Mathematics and Statistics
M.S. in Applied Mathematics
University of Missouri at Rolla
iv, 25 pages
© 1964 William Lawrence May, All rights reserved.
Thesis - Open Access
Print OCLC #
Link to Catalog Record
May, William Lawrence, "A study on estimating parameters restricted by linear inequalities" (1964). Masters Theses. 5647.