⚠️Fisher information may be undefined (infinite) when the integral the covariance of the score diverge
📙 Slide describes a simple example using location model with finite support of its standard density.
➡️regular vs non-regular statistical models
👉mdpi.com
📙 Slide describes a simple example using location model with finite support of its standard density.
➡️regular vs non-regular statistical models
👉mdpi.com
It is important to have distribution support not parameterized by parameters like exponential families to have well-defined Fisher information.
Loading suggestions...