On Asymptotic Properties of a Maximum Likelihood Estimator of Stochastically Ordered Distribution Function Oh, Myongsik;
Kiefer (1961) studied asymptotic behavior of empirical distribution using the law of the iterated logarithm. Robertson and Wright (1974a) discussed whether this type of result would hold for a maximum likelihood estimator of a stochastically ordered distribution function; however, we show that this cannot be achieved. We provide only a partial answer to this problem. The result is applicable to both estimation and testing problems under the restriction of stochastic ordering.
Law of the iterated logarithm;maximum likelihood estimation;stochastically ordered distribution function;
Billingsley, P. (1986). Probability and Measure, 2nd ed., Wiley, New York.
Kiefer, J. (1961). On large deviation of the empirical distribution function of vector random variable and a law of the iterative logarithm, Pacific Journal of Mathematics, 11, 649-660.
Lee, C. I. C. (1987). Maximum Likelihood Estimates for Stochastically Ordered Multinomial Populations with Fixed and Random Zeros, Foundation of Statistical Inference, (I. B. MacNeill and G. J. Umphrey eds.) 189-197.
Robertson, T. and Wright, F. T. (1974a). On the maximum likelihood estimation of stochastically ordered random variables, The Annals of Statistics, 2, 528-534.
Robertson, T. and Wright, F. T. (1974b). A norm reducing property for isotonized cauchy mean value functions, The Annals of Statistics, 2, 1302-1307.
Robertson, T. and Wright, F. T. (1981). Likelihood ratio tests for and against a stochastic ordering between multinomial populations, The Annals of Statistics, 9, 1248-1257.
Robertson, T., Wright, F. T. and Dykstra, R. L. (1988). Order Restricted Statistical Inference, Wiley, Chichester.