Nonparametric estimation of regression functions with both categorical and continuous data

By Jeffrey S. Racine, Qi Li. 2004

384 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/2aa2e117b03cb7c081ee619b2143df3ec0769298

How researchers see this paper

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

In Statistics and Probability

Citations in the corpus

The Impact of Homework on Student Achievement (2008, methodology): “To proceed, we utilize Li-Racine Generalized Kernel Estimation (Li and Racine, 2004 and Racine and Li, 2004) and express the test score equation as TSi = θ (xi ) + ei , i = 1, . . . , N , (3.2) where θ (·) is the unknown smooth educational production function, ei is an additive error term and N is…”
Jackknife model averaging (2012, methodology): “We also consider two nonparametric kernel estimators, namely, the local linear (NPll) and local constant (NPlc) estimators with datadriven bandwidths modified to handle categorical and continuous data; see Li and Racine (2004) and Racine and Li (2004) for details.”
The Relationship between Stock Returns and Volatility in International Stock Markets (2005, background): “In nonparametric kernel estimations, Hurvich et al. (1998), Li and Racine (2004), and Racine and Li (2004) show that an improved (a modified) AIC criterion performs well in selecting smoothing parameters.”
How robust is the evidence on the effects of college quality? Evidence from matching (2004, methodology): “We use a least squares leave-one-out validation mechanism to choose among the nearest neighbor, Gaussian kernel, and Epanechnikov kernel estimators and to pick the bandwidth for the two kernel estimators; see Racine and Li (2004) and Pagan and Ullah (1999) for discussions of leave-one-out validati”
How to improve the performances of DEA/FDH estimators in the presence of noise? (2007, positive): “Henderson and Simar (2005) suggest an alternative model where the number of time periods T is fixed and consistency requires only n !1 . Henderson and Simar obtain fully nonparametric estimates of the frontier and the time-variant technical efficiencies by exploiting recent advances in kernel regres”
Estimation of Quantile Treatment Effects with Stata (2010, methodology): “At a higher level, locreg implements nonparametric estimation with both categorical and continuous regressors as suggested by Racine and Li (2004).”
Including Covariates in the Regression Discontinuity Design (2018, methodology): “Product kernel functions also have the advantage that one can easily incorporate discrete X in the spirit of Racine and Li (2004).”
Identification and estimation of semiparametric two‐step models (2016, methodology): “For our proposed estimator, estimator III, in the first stage we nonparametrically regress the endogenous covariate log(Income) on State, Edu, Size, and Age using generalized kernels as suggested in Racine and Li (2004) to handle the combination of discrete and continuous regressors, with smoothing ”
Volatility jumps: The role of geopolitical risks (2018, methodology): “1 For each quantile, we determine the bandwidth using the leave-one-out least-squares cross validation method of Racine and Li (2004) and Li and Racine (2004).”
Threshold Regression With a Threshold Boundary (2020, methodology): “When x and/or z contain discrete variables, we can extend the method in Racine and Li (2004) from nonparametric regression to density estimation.”
Nonparametric Estimation of Regression Functions in the Presence of Irrelevant Regressors (2007, background): “More recently, Racine and Li (2004) have considered the nonparametric estimation of regression functions”
A consistent model specification test with mixed discrete and continuous data (2007, methodology): “More specifically, defining h0s = a 0 sn −1/(4+q) and λ0s = b 0 sn −2/(4+q), where a0s is a positive constant and b0s is a non-negative constant, 4 Racine and Li (2004) (assuming all regressors are relevant) show that (ĥs − h0s)/h0s = Op(n− /(4+q)) for s = 1, . . . , q, and λ̂s − λ0s = Op(n−δ) for…”