| Qi Li, Texas A & M University |
| Nonparametric Estimation with Categorical and Continuous Data |
| Session: C-12-20 Wednesday 16 August 2000 by Li, Qi |
| In this paper we propose a method for nonparametric regression which admits continuous and categorical data in a natural manner using the method of kernels. The strength of the method lies in its ability to model situations involving complex dependence between the variable predicted and mixed discrete and continuous explanatory data types. The method extends earlier nonparametric estimation method for mupltivariate categorical data to the case of mixed discrete and continuous data models. A data-driven method of bandwidth selection is proposed, and consistency and rate of convergence of the proposed estimator are established. Simulations suggest that the new estimator performs much better than the conventional nonparametric estimator. Two empirical applications are considered, one involving the estimation of earnings profiles and the other involving the estimation of Engle curves. These application demonstrate the ability of the technique to detect structure in mixed categorical and continuous data which sometimes remain undetected by traditional parametric approaches, and show that the new estimator are superior in out-of-sample prediction to the conventional nonparametric estimator as well as commonly used parametric estimators. |