7

I have created two fractions (see example below), but the denominator is a little too close to the division bar. Can I change this somehow?

\documentclass{article}
\usepackage{amsmath}

\begin{document}
    Combining two Gaussians with mean $\mu_1, \mu_2$ and variance $\sigma_1^2, \sigma_2^2$ yields a new Gaussian with mean $\mu = \frac{\sigma_2^2 \mu_1 + \sigma_1^2 \mu_2}{\sigma_1^2 + \sigma_2^2}$ and variance $\sigma^2 = \frac{1}{\frac{1}{\sigma_1^2} + \frac{1}{\sigma_2^2}}$
\end{document}
2
  • 3
    Use \dfrac instead of \frac.
    – user156344
    Mar 8, 2019 at 11:02
  • You could add a \vphantom command (for examle \vphantom{A^{A^A}}) in the beginning of each problematic dominator, and keep the style as inline math.
    – koleygr
    Mar 8, 2019 at 11:06

3 Answers 3

9

You have two main options:

  • Switch from \frac{...}{...}-notation to inline-fraction notation

  • Switch to display math to typeset the formulas for \mu and \sigma^2.

enter image description here

\documentclass{article}
\usepackage{amsmath} % for "\text" macro
\begin{document}

\noindent
1. OP's original version:

Combining two Gaussians with mean $\mu_1, \mu_2$ and variance $\sigma_1^2, \sigma_2^2$ yields a new Gaussian with mean $\mu = \frac{\sigma_2^2 \mu_1 + \sigma_1^2 \mu_2}{\sigma_1^2 + \sigma_2^2}$ and variance $\sigma^2 = \frac{1}{\frac{1}{\sigma_1^2} + \frac{1}{\sigma_2^2}}$.

\medskip\noindent
2. Partial switch to inline-math notation

Combining two Gaussians with mean $\mu_1, \mu_2$ and variance 
$\sigma_1^2, \sigma_2^2$ yields a new Gaussian with mean 
$\mu = \frac{\sigma_2^2 \mu_1 + \sigma_1^2 \mu_2}{\sigma_1^2 + \sigma_2^2}$ 
and variance $\sigma^2 = \frac{1}{1/\sigma_1^2 + 1/\sigma_2^2}$.

\medskip\noindent
3. Full switch to inline math notation

Combining two Gaussians with means $\mu_1$ and $\mu_2$ and 
variances $\sigma_1^2$ and $\sigma_2^2$ yields a new Gaussian 
with mean $\mu = (\sigma_2^2 \mu_1 + \sigma_1^2 \mu_2)/(\sigma_1^2 + 
\sigma_2^2)$ and variance $\sigma^2 = 1/(1/\sigma_1^2 + 1/\sigma_2^2)$.

\medskip\noindent
4. Switch to display math

Combining two Gaussians with means $\mu_1$ and $\mu_2$ and 
variances $\sigma_1^2$ and $\sigma_2^2$ yields a new Gaussian 
with mean $\mu$ and variance $\sigma^2$ given by
\[
\mu=\frac{\sigma_2^2 \mu_1 + \sigma_1^2 \mu_2}{\sigma_1^2 + \sigma_2^2} 
\quad\text{and}\quad 
\sigma^2 = \frac{1}{1/\sigma_1^2 + 1/\sigma_2^2}\,.
\]

\end{document} 
4
  • \everymath{\displaystyle} would be an additional option (Just in case that the document can afford such thing.) Of course I would not use it often, but have used in special cases that looked ok. (+1)
    – koleygr
    Mar 8, 2019 at 11:23
  • 1
    @koleygr - Unless the paragraph in question is double-spaced (and hence already damaged beyond repair, typographically speaking), typesetting a \dfrac expression in running text must surely be a high crime against all known forms of decent typography. :-) If you doubt this claim, just look at the outputs of Sebastiano's and Steven's solutions...
    – Mico
    Mar 8, 2019 at 11:27
  • 1
    Thanks for the info, I also didn't like these kind of solutions but just added the comment in order to propose a more general answer for every case. But of course you are right!
    – koleygr
    Mar 8, 2019 at 11:30
  • 1
    went for the last option. Thx, Mico!
    – Luk
    Mar 8, 2019 at 11:38
5

Here, I preserve the fraction in its native \textstyle, but add a (default) 1pt buffer above and below the numerator and denominator of each fraction, which can be changed with an optional argument. I call it \qfrac[]{}{}. The MWE shows before and after.

\documentclass{article}
\usepackage{stackengine,scalerel}
\stackMath
\newcommand\qfrac[3][1pt]{\frac{%
  \ThisStyle{\addstackgap[#1]{\SavedStyle#2}}}{%
  \ThisStyle{\addstackgap[#1]{\SavedStyle#3}}%
}}
\usepackage{amsmath}

\begin{document}
    Combining two Gaussians with mean $\mu_1, \mu_2$ and variance $\sigma_1^2, 
  \sigma_2^2$ yields a new Gaussian with mean $\mu = \frac{\sigma_2^2 \mu_1 + 
  \sigma_1^2 \mu_2}{\sigma_1^2 + \sigma_2^2}$ and variance $\sigma^2 = 
  \frac{1}{\frac{1}{\sigma_1^2} + \frac{1}{\sigma_2^2}}$

   Combining two Gaussians with mean $\mu_1, \mu_2$ and variance $\sigma_1^2, 
  \sigma_2^2$ yields a new Gaussian with mean $\mu = \qfrac{\sigma_2^2 \mu_1 + 
  \sigma_1^2 \mu_2}{\sigma_1^2 + \sigma_2^2}$ and variance $\sigma^2 = 
  \qfrac[.5pt]{1}{\qfrac{1}{\sigma_1^2} + \qfrac{1}{\sigma_2^2}}$
\end{document}

enter image description here

2
  • 1
    (+1)... I would add your "1pt" inside an extra optional argument of the command.
    – koleygr
    Mar 8, 2019 at 11:18
  • Nice answer... Straightforward to the OP's problem. Welcome.
    – koleygr
    Mar 8, 2019 at 11:21
0

Alternatively, use \raisebox :

\documentclass{article}
\usepackage{amsmath}
\begin{document}
   Combining two Gaussians with mean $\mu_1, \mu_2$ and variance $\sigma_1^2, \sigma_2^2$ yields a new Gaussian with mean $\mu = \frac{\raisebox{.2in}{$\sigma_2^2 \mu_1 + \sigma_1^2 \mu_2$}}{\raisebox{-.2in}{$\sigma_1^2 + \sigma_2^2$}}$ and variance $\sigma^2 = \frac{\raisebox{.2in}{$1$}}{\raisebox{-.2in}{$\frac{1}{\sigma_1^2} + \frac{1}{\sigma_2^2}$}}$
\end{document}

result

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .