DOC: Fix mathematical notation in `conditional_logprob` docstrings

[larryshamalama]

Description

Nitpick issue here

• $\Sigma$ is conventionally used for covariance matrices, so I prefer using $\sigma$ since the example is univariate.

pymc/pymc/logprob/basic.py

Lines 433 to 434 in 7b08fc1

	\Sigma^2 \sim& \operatorname{InvGamma}(0.5, 0.5) \\
	Y \sim& \operatorname{N}(0, \Sigma)

• For continuous-valued random variables, I would prefer not writing $P(Y = y)$ because, in principle, it is always zero.
  

  pymc/pymc/logprob/basic.py

  Lines 448 to 449 in 7b08fc1

  	\log p(Y = y, \Sigma^2 = \sigma^2) =
  	\log p(Y = y \mid \Sigma^2 = \sigma^2) + \log p(\Sigma^2 = \sigma^2)

What I would recommend is perhaps to use the random variables as indices to the log-density and valued variables as arguments to delineate their differences while addressing the point above: $\log p_{Y, \sigma^2}(y, s^2) = \log p_{Y \mid \sigma^2}(y \mid \sigma^2 = s^2) + \log p_{\sigma^2}(s^2)$. To me, this is better, but maybe not for everyone...

The last instance where change may be needed is:

pymc/pymc/logprob/basic.py

Line 438 in 7b08fc1

conditional log-probability :math:`\log p(Y = y \mid \Sigma^2)`, with a stochastic

[ricardoV94]

I think I made it capital sigma to contrast with lower case, but using s is also fine.

No objections to the second form.

[amyoshino]

I’m working on this next :)