Abstract
Multiple regression is employed in cost estimation with prudence because of the frequently encountered problem of multicollinearity. This existence of multicollinearity causes a loss in the precision of the estimates of the regression coefficients and their related standard errors. Ridge regression is a biased regression estimation procedure that results in a substantial reduction in the total variance of a system while only slightly increasing the bias. In this paper it is demonstrated how problems resulting from the existence of multicollinearity in cost estimation can be overcome by employing ridge regression.
Original language | English (US) |
---|---|
Pages | 19-21 |
Number of pages | 3 |
State | Published - 1979 |
Event | Unknown conference - New Orleans, LA, USA Duration: Nov 19 1979 → Nov 21 1979 |
Other
Other | Unknown conference |
---|---|
City | New Orleans, LA, USA |
Period | 11/19/79 → 11/21/79 |
All Science Journal Classification (ASJC) codes
- General Engineering