How to plot a marginal histogram correctly? Why is my plot different to the one in a textbook?

Question

Chapter 1 of "Machine Learning - A Probabilistic Perspective" by Kevin Patrick Murphy gives this figure (fig_1),

and says

this is a pairwise scatter plot on iris dataset. The diagonal plots the marginal histograms of the 4 features.

This post, gives this figure (fig_2) to illustrate "marginal histograms".

I am trying to plot a histogram by using the approach that plotted out fig_2, with first feature of iris dataset, that is, sepal length.

Here is the code,

import matplotlib.pyplot as plt
from sklearn import datasets
import seaborn as sns; 
sns.set(style="white", color_codes=True)
iris = datasets.load_iris()
_ = plt.hist(iris.data[:,0], bins = 11)

which plotted out this figure (fig_3.)

Which looks different to the top left in fig_1.

I've searched whole site and got this and this, none of them gives explanation what a marginal histograms is.

How to plot a marginal histogram correctly? Why is my plot different to the one in a textbook?

@Glen_b I've updated the OP, would you please take a loot at it? — JJJohn, Sep 26 '19 at 06:14
I put it on hold to protect it while you edited. You sound like you're done, so I've reopened. — Glen_b, Sep 26 '19 at 06:26
A marginal histogram is a histogram of the data. That's all there is to it. Whether and how you can plot them is a matter of consulting your software documentation. — whuber, Sep 27 '19 at 15:05
"Marginal" is a standard term for focusing on a single univariate component of a multivariate distribution. Thus, "marginal histograms" are histograms of the individual variables. — whuber, Sep 27 '19 at 17:50

score 0 · Answer 1 · answered Jul 01 '23 at 12:07

0

There are subtle differences between the two histograms, but we cannot infer much based on this and it doesn't necessarily mean that either is incorrect. Histograms based on the same data can differ if the bin widths or intervals are different - it's quite possible that has happened here.

answered Jul 01 '23 at 12:07

mkt

18,245
11
73
172

In particular, the text-book histograms have 12 divisions, the OP's has only 11. – Ray Butterworth Jul 01 '23 at 12:29

How to plot a marginal histogram correctly? Why is my plot different to the one in a textbook?

1 Answers1