论文信息 - ASYMPTOTIC PARAMETRIC TESTS OF NONLINEAR HYPOTHESES

ASYMPTOTIC PARAMETRIC TESTS OF NONLINEAR HYPOTHESES

Abatnet The parametric asymptotic test theory for composite hypotheses is presented in a rather general set-up: (i) the composite hypotheses are of a truly nonlinear form (ii) the norming matrices are not necessarüy diagonal (iü) the underlying stochastic model forms a 'non-ergodic model' (iv) the 'objective' or 'estimation' function approach is chosen. AUowing (i) and (ii) simultaneously enforces a condition on the compatibility of the norming matrices and the form of the hypothesis. The familiar asymptotic x^ " distributions of the likelihood ratio-, the scoreand the Wald-statistic are derived. An example involving a parametric trend function iUustrates some aspects of the approach.

H. Pruscha

[1] David Lindley,et al. Statistical inference for Markov processes , 1961 .

[2] W. E. Lever,et al. THE LIMITING DISTRIBUTION OF THE LIKELIHOOD RATIO STATISTIC UNDER A CLASS OF LOCAL ALTERNATIVES. , 1967 .

[3] H. Koul,et al. Asymptotic tests of composite hypotheses for non-ergodic type stochastic processes , 1979 .

[4] Arjun K. Gupta. The Theory of Linear Models and Multivariate Analysis , 1981 .

[5] L. Fahrmeir. Asymptotic testing theory for generalized linear models , 1987 .

[6] A. Gallant,et al. Nonlinear Statistical Models , 1988 .

[7] L. Fahrmeir. A note on asymptotic testing theory for nonhomogeneous observations , 1988 .

[8] H. Pruscha. Angewandte Methoden der Mathematischen Statistik , 1996 .