Faculty salary equity: Issues in regression model selection

1993 ◽  
Vol 34 (1) ◽  
pp. 107-126 ◽  
Author(s):  
Nelle Moore
Keyword(s):  
Model Selection ◽  
Regression Model ◽  
Faculty Salary ◽  
Equity Issues ◽  
Salary Equity

Faculty salary equity: Issues and options

1994 ◽  
Vol 35 (1) ◽  
pp. 1-19 ◽  
Author(s):  
Julie K. Snyder ◽  
Patricia B. Hyer ◽  
Gerald W. McLaughlin
Keyword(s):  
Faculty Salary ◽  
Equity Issues ◽  
Salary Equity

On stochastic linear regression model selection

2019 ◽  
Author(s):  
J. Peter Praveen ◽  
B. Mahaboob ◽  
Ranadheer Donthi ◽  
S. Vijay Prasad ◽  
B. Venkateswarlu
Keyword(s):  
Model Selection ◽  
Linear Regression ◽  
Regression Model ◽  
Linear Regression Model

scholarly journals Role of regression model selection and station distribution on the estimation of oceanic anthropogenic carbon change by eMLR

2012 ◽  
Vol 9 (10) ◽  
pp. 14589-14638 ◽  
Author(s):  
Y. Plancherel ◽  
K. B. Rodgers ◽  
R. M. Key ◽  
A. R. Jacobson ◽  
J. L. Sarmiento
Keyword(s):  
Model Selection ◽  
Regression Model ◽  
North Atlantic ◽  
Data Sets ◽  
Linear Regression Models ◽  
The North ◽  
Basin Scale ◽  
Anthropogenic Carbon ◽  
Hydrographic Features ◽  
Carbon Change

Abstract. Differencing predictions of linear regression models generated from hydrographic data collected at different times (the eMLR method) was proposed as a means of quantifying the dominant patterns of change in oceanic anthropogenic carbon in the context of sparse data sets subject to natural variability. The ability of eMLR to recover the anthropogenic carbon signal in the North Atlantic was tested using a global circulation and biogeochemistry model. Basin-scale applications of eMLR on horizontal layers can estimate the change in anthropogenic carbon inventory with an accuracy typically better than 10%. Regression model selection influences the distribution of the recovered anthropogenic carbon change signal. The systematic use of statistically optimum regression formulae does not produce the best estimates of anthropogenic carbon change if the distribution of the station locations emphasizes hydrographic features differently in time. Additional factors, such as a balanced station distribution and vertical continuity of the regression formulae should be considered to guide model selection. Accurate results are obtained when multiple formulae are used throughout the water column. Different formulae can yield results of similar quality. The fact that good results are obtained in the hydrographically complex North Atlantic suggests that eMLR can produce accurate estimates in other basins.

Sign in / Sign up

Export Citation Format

Share Document