Hostname: page-component-cd9895bd7-jkksz Total loading time: 0 Render date: 2024-12-27T18:49:46.489Z Has data issue: false hasContentIssue false
  • English
  • Français

Testing the Goodness of Fit of a Parametric Density Function by Kernel Method

Published online by Cambridge University Press:  11 February 2009

Yanqin Fan
Get access Rights & Permissions [Opens in a new window]

Abstract

Let F denote a distribution function defined on the probability space (Ω,,P), which is absolutely continuous with respect to the Lebesgue measure in Rd with probability density function f. Let f0(·,β) be a parametric density function that depends on an unknown p × 1 vector β. In this paper, we consider tests of the goodness-of-fit of f0(·,β) for f(·) for some β based on (i) the integrated squared difference between a kernel estimate of f(·) and the quasimaximum likelihood estimate of f0(·,β) denoted by In and (ii) the integrated squared difference between a kernel estimate of f(·) and the corresponding kernel smoothed estimate of f0(·, β) denoted by Jn. It is shown in this paper that the amount of smoothing applied to the data in constructing the kernel estimate of f(·) determines the form of the test statistic based on In. For each test developed, we also examine its asymptotic properties including consistency and the local power property. In particular, we show that tests developed in this paper, except the first one, are more powerful than the Kolmogorov-Smirnov test under the sequence of local alternatives introduced in Rosenblatt [12], although they are less powerful than the Kolmogorov-Smirnov test under the sequence of Pitman alternatives. A small simulation study is carried out to examine the finite sample performance of one of these tests.

Type
Research Article
Information
Econometric Theory , Volume 10 , Issue 2 , June 1994 , pp. 316 - 356
Copyright
Copyright © Cambridge University Press 1994

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1.Ahmad, I.A. & Van Belle, G.. Measuring affinity of distributions. In Proschan, F. & Serfling, R.J. (eds.), Reliability and Biometry: Statistical Analysis of Lifelength, pp. 671–668. Philadelphia: Society for Industrial and Applied Mathematics, 1974.Google Scholar
2. Bickel, P. J. & Rosenblatt, M. On some global measures of the deviations of density function estimates. Annals of Statistics 1 (1973): 1071–1095.10.1214/aos/1176342558CrossRefGoogle Scholar
3. Fan, Y. & Gencay, R. Hypotheses testing based on modified nonparametric estimation of an affinity measure between two distributions. Journal of Nonparametric Statistics 2 (1993): 389–403.10.1080/10485259308832567Google Scholar
4. Ghosh, B.K. & Huang, W. The power and optimal kernel of the Bickel-Rosenblatt test for goodness of fit. Annals of Statistics 19 (1991): 999–1009.10.1214/aos/1176348133Google Scholar
5. Hall, P. Central limit theorem for integrated square error of multivariate nonparametric density estimators. Journal of Multivariate Analysis 14 (1984): 1–16.10.1016/0047-259X(84)90044-7CrossRefGoogle Scholar
6. Hall, P. & Marron, J.S.. Estimation of integrated squared density derivatives. Statistics & Probability Letters 6 (1987): 109–115.10.1016/0167-7152(87)90083-6CrossRefGoogle Scholar
7. Härdle, W. & Mammen, E. Comparing nonparametric versus parametric regression fits. Annals of Statistics 21 (1993): 1926–1947.10.1214/aos/1176349403Google Scholar
8. Prakasa, Rao B.L.S. Nonparametric Functional Estimation. New York: Academic Press, 1983.Google Scholar
9. Ramberg, J.S. & Schmeiser, B.W.. An approximate method for generating asymmetric random variables. Communications of the Association for Computing Machinery 17 (1974): 78–82.Google Scholar
10. Randies, R.H., Fligner, M.A., Policello, G.E. II & Wolfe, D.A.. An asymptotically distribution-free test for symmetry versus asymmetry. Journal of the American Statistical Association 75 (1980): 168–172.CrossRefGoogle Scholar
11. Robinson, P.M. Consistent nonparametric entropy-based testing. The Review of Economic Studies 58 (1991): 437–453.10.2307/2298005CrossRefGoogle Scholar
12. Rosenblatt, M.A. Quadratic measure of deviation of two-dimensional density estimates and a test of independence. Annals of Statistics 3 (1975): 1–14.10.1214/aos/1176342996CrossRefGoogle Scholar
13. White, H. Maximum likelihood estimation of misspecified models. Econometrica 50 (1982): 1–25.10.2307/1912526CrossRefGoogle Scholar