normal-corner 0.0.1

Analytical corner plot for Gaussian distributions

normal_corner

This code produces a corner plot for analytical multi-dimensional Gaussian distribution, using covariance matrix and mean matrix. It also allows us to plot another distribution, with reduced dimensionality, on top. I.e. a distribution for a case where we fixed one variable.

See demo.py for examples.

Installation

Option 1:

• pip install normal_corner

Option 2:

• python setup.py install in cloned directory

Documentation

The main component is function normal_corner inside a normal_corner package. Below is a description of inputs and outputs.

Output:

• Matplotlib figure object with a corner plot

Main input:

• covm : covariance matrix, numpy array, NxN.
• mean : mean matrix, numpy array, 1xN.
• varlabels : labels for plotting, 1xN, list of str in LaTex format, between ($$).

Input for a second distribution on top:

• fixedvarindex : index of variable that we do not use (fix), int, starting from 0. If not None, covm2 and mean2 must not be None.
• fixedvarvalue : value of fixed variable, float. Leave None not to plot fixed value.
• newcov and newmean : new covariance and mean matrices, same format, as above.

Optional input:

• scale_factor : scale factor for plotting area, float, in sigma.
• diagnostic : an option to print out some diagnostic messages, bool.
• color : color for a main Normal distribution, str.
• color2 : color for a secondary Normal distribution, with reduced dimensionality, str.