How to add correct labels for Seaborn Confusion Matrix

Question

I have plotted my data into a confusion matrix using seaborn but I ran into a problem. The problem is that it is only showing numbers from 0 to 11, on both axes, because I have 12 different labels.

My code looks as follows:

cf_matrix = confusion_matrix(y_test, y_pred)
fig, ax = plt.subplots(figsize=(15,10)) 
sns.heatmap(cf_matrix, linewidths=1, annot=True, ax=ax, fmt='g')

Here you can see my confusion matrix:

I am getting the confusion matrix as I should. The only problem is the names of the labels which are not shown. I have searched quite a while over the Internet and with no luck. Are there any parameters which can attach the labels or how can this be done?

Answer 1

When you factorize your categories, you should have retained the levels, so you can use that in conjunction with pd.crosstab instead of confusion_matrix to plot. Using iris as example:

import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
from sklearn.ensemble import RandomForestClassifier
from sklearn.datasets import make_classification
from sklearn.metrics import classification_report, confusion_matrix

df = pd.read_csv("http://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data",
                 header=None,names=["s.wid","s.len","p.wid","p.len","species"])
X = df.iloc[:,:4]
y,levels = pd.factorize(df['species'])

At this part, you get the labels y in [0,..1,..2] and levels as the original labels to which 0,1,2 corresponds to:

Index(['Iris-setosa', 'Iris-versicolor', 'Iris-virginica'], dtype='object')

So we fit and do like what you have:

clf = RandomForestClassifier(max_depth=2, random_state=0)
clf.fit(X,y)
y_pred = clf.predict(X)
print(classification_report(y,y_pred,target_names=levels))

And a confusion matrix with 0,1,2:

cf_matrix = confusion_matrix(y, y_pred)
sns.heatmap(cf_matrix, linewidths=1, annot=True, fmt='g')

We go back and use the levels:

cf_matrix = pd.crosstab(levels[y],levels[y_pred])
fig, ax = plt.subplots(figsize=(5,5))
sns.heatmap(cf_matrix, linewidths=1, annot=True, ax=ax, fmt='g')