|
Identifiability of a Linear Relation between Variables Which Are Subject to Error
Olav Reiersol
Abstract
The problem of the identifiability of a linear relation between variables subject to error is examined in two different two-variable models in which the errors are independent of the "true" variables. In one model the errors are specified to be jointly normally distributed, and in the other to be stochastically independent. In both cases necessary and sufficient conditions for identifiability are found. A summary of previous results is included.
|
