I have the following code:
fig, ax = plt.subplots(figsize=(15, 15))
plot_confusion_matrix(clf, X_test, y_test, ax=ax)
print(pd.Series(y_test).value_counts().sort_index())
The output is shown below. What I don't understand is why the unique values of y_test do not match the True Label in the confusion matrix. For example, 46 to 50 are missing, as well as e.g. 20. Those values do appear in y_test but not in the confusion matrix. Why could this be? I referenced the manual but can't understand why it happens.
10 1
20 2
23 1
26 1
28 2
29 1
30 19
31 3
32 13
33 10
34 31
35 39
36 99
37 131
38 287
39 321
40 499
41 501
42 644
43 438
44 382
45 214
46 113
47 45
48 29
49 3
50 21
dtype: int64