Hi,

The docstring contains a variable normalize.
However, this is not a legal parameter to pass to the function make_confusion_matrix().
Thank you for openly publishing this function :)

Hello,

thanks a lot for the amazing code. Yesterday it was working fine. today it is giving me this error

AttributeError: 'QuadMesh' object has no property 'fontize'
strange I didn't change anything

thanks and regards,

Hi I found your blog post https://medium.com/@dtuk81/confusion-matrix-visualization-fc31e3f30fea helpful and visited the github page.

The project does not have a license file. Since the blog says "Feel free to copy my code or to make suggestions on how to update the functionality" should I treat this as a MIT/Apache licensed project by default?

This code works great but the y-tick marks and annotations are misaligned rather than centered. I have changed my version of matplotlib as it seemed to be a known version issue and even used the "plt.yticks([0.5,1.5]" trick but still the labels are not centered.

Snippet of Code:
# MAKE THE HEATMAP VISUALIZATION
plt.figure(figsize=figsize)
sns.heatmap(cf,annot=box_labels,fmt="",cmap=cmap,cbar=cbar,xticklabels=categories,yticklabels=categories)
if xyplotlabels:
plt.ylabel('True', fontsize=16)
plt.xlabel('Predicted' + stats_text, fontsize=16)
plt.xticks([0.5,1.5], fontsize=16)
plt.yticks([0.5,1.5], fontsize=16, ha="center", va="bottom")
else:
plt.xlabel(stats_text)

if title:
    plt.title(title)

make_confusion_matrix(cm_rf)

Hello,

Thank you for the function. I ran into 1 issue while saving the chart into png. The error was "'NoneType' object has no attribute 'savefig'". Do you know what would have cause the object to be classify as NoneType?

The percentage calculation was incorrect (python 3).
The percentage was based on the total sum of cf, but should be based on the column sums.
I am using the following hack but I am sure the code can be more elegant.

if percent:
    tmp = cf/np.sum(cf,axis=1).reshape(-1,1)
    group_percentages = ["{0:.1%}".format(value) for value in tmp.flatten()]

Hello what a good graph

I have a question, what is the difference between Accurancy and precision