Some Comments onCp

By C. L. Mallows. 1973

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Efficiency for Regularization Parameter Selection in Penalized Likelihood Estimation of Misspecified Models (2013, background): “First, we show that the feasible version of GIC2, which corresponds to the wellknown Cp measure (Mallows 1973), is in fact efficient in the nontrue model world.”
Bayesian CART Model Search (1998, methodology): “As illustrated in Section 7, a useful tool in this regard is a plot of the largest observed values of p(Y jX;T ) against the number of terminal nodes of T , an analogue of the Cp plot (Mallows 1973).”
Generalized Least Squares Model Averaging (2016, methodology): “Numerous model selection techniques, such as the Akaike information criterion (Akaike, 1973), Mallows’ Cp criterion (Mallows, 1973), and the Bayesian information criterion (Schwarz, 1978), have been proposed, and many researchers have investigated their properties.”
Evaluation of infiltration models with different numbers of fitting parameters in different soil texture classes (2014, methodology): “The F-statistic (Green & Caroll 1978) was used to compare model goodness-of-fit to soilmoisture characteristic data by Vereecken et al. (1989); Buchan et al. (1993) applied the F-statistic and the Cp statistic of Mallows (1973) to find the best-fit particle-size distribution model.”
Cross-validation: what does it estimate and how well does it do it? (2021, methodology): “For the linear model with OLS and squared error loss, this approach specializes to the well-known Mallows C p (Mallows 1973; Akaike 1974) estimate of prediction error: We first consider the estimation of prediction error with Mallows C p , showing that (like cross-validation) it is a worse…”
The Practical Implementation of Bayesian Model Selection (2001, background): “For example F = 2 yields Cp (Mallows 1973) and, approximately, AIC (Akaike 1973), F = log n yields BIC (Schwarz 1978) and F = 2 log p yields RIC (Donoho and Johnstone 1994, Foster and George 1994).”
A Model-Averaging Approach for High-Dimensional Regression (2014, methodology): “…Hoeting 1997; Hoeting et al. 1999; Fernandez, Ley, and Steel 2001), the Mallows Cp model averaging (Hansen 2007; Wan, Zhang, and Zou 2010; see also Mallows 1973), Bayesian model averaging using predictive measures (Eklund and Karlsson 2007), jackknife model averaging (Hansen and Racine 2012), and…”
Methods and Criteria for Model Selection (2004, background): “A related criterion is the Cp statistic (Mallows 1973), Cp D RSSp = O¾ 2 C .2p ¡ n/; (27) where RSSp represents the residual sum of squares for a model with p terms and O¾ 2 denotes the estimate of the error variance based on the full model.”
Estimation and Accuracy after Model Selection (2013, methodology): ““Cubic” was selected by the Cp criterion (Mallows 1973), as described in Section 2.”
Model selection for ecologists: the worldviews of AIC and BIC. (2014, methodology): “…consistent estimators, namely BIC, Hannan and Quinn information (Hannan and Quinn 1979), and GM (Geweke and Meese 1981), and efficient estimators, namely AIC, Mallows’ Cp (Mallows 1973), predicted residual sum of squares (PRESS; Allen 1974), Akaike’s FPE (Akaike 1969), and cross validation.”
Model Selection and the Principle of Minimum Description Length (2001, background): “It can be shown that for a ﬁxed dimension k , gMDL will select the lowest RSS model, a property shared by BIC and AIC as well as other selection criteria like Mallows’ C p (Mallows, 1973).”
A group bridge approach for variable selection (2009, methodology): “Traditional approaches to variable selection include C p (Mallows, 1973), AIC (Akaike, 1973), and BIC (Schwarz, 1978).”
Default Bayes Factors for Model Selection in Regression (2012, methodology): “Some methods favor models that provide for better out-of-sample prediction, including the Akaike information criterion (AIC; Akaike, 1974) and Mallows’s C p (Mallows, 1973).”
Empirical Bayes vs. Fully Bayes Variable Selection (2008, background): “Such criteria include AIC (Akaike, 1973) and Cp (Mallows, 1973) when F = 2; BIC (Schwarz, 1978) when F = log n; and RIC (Foster and George, 1994) when F = 2 log p.”