ABSTRACT
Consider a family of probability spaces (fi, A, Pq) indexed by a parameter 9 G 0 C R and let X i , . . . , X n be a random sample from some specific (but unknown) distribution Pq in this family. In this context, the objective of p a r a m e tr ic in fe re n c e is to propose and evaluate methods for choosing appropriate statistics Tn = Tn( X i , . . . , X n) to e s t im a te 9 (i.e., to guess the true value of 9) or to te s t h y p o th e se s about 9 (i.e., to decide whether 9 G ©o C 0 or not); in the first case Tn is called an e s t im a to r of 9 (and is usually denoted as 9n ) and in the second case a te s t s ta t i s t ic .
