This is one of my first questions so sorry in advance for any rules that I break.

Is there any mainstream convention for the symbol of the precision and precision matrix?

I have seen the following ones:

  • In Information theory, inference and learning algorithms by MacKay:

$$ \text{Precision: }\tau = 1 / \sigma^2$$ $$ \text{Precision matrix: }\mathbf{A} = \boldsymbol{\Sigma}^{-1} $$

  • In Pattern recognition and machine learning by Bishop:

$$ \text{Precision: }\beta = 1 / \sigma^2$$ $$ \text{Precision matrix: }\boldsymbol{\Lambda} = \boldsymbol{\Sigma}^{-1} $$

Is any of them preferred over the others?


1 Answer 1


I change a lot but in the Gaussian Process literature, it's usually $K$ (for kernel matrix) or $\Sigma$. Precision matrices also change a lot but as you say $A$ and $\Lambda$ are pretty common. As long as you clearly introduce the term though you can use whichever convention you want.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.