Why is my confusion matrix like this, when i tried all the tests and I got reasonable scores?

I am using the random forest classification from sklearn and I am getting decent results in everything except the confusion matrix here are the codes and results

The label distribution for the training and testing

The size of the train set

The model

Scores for the training model

Here is the issue

This is not what i was expecting especially since that the amount of the training was only 1/3 of what it should have been i have 677k in the training dataset, but in the confusion matrix it's only doing all the label 0's.

The model:

import time
# Record the starting time
start_time = time.time()

# Random Forest classifier
rf = RandomForestClassifier()

# Define the parameter grid
rf_param_grid = {'n_estimators': [45], 'criterion': ['entropy'], 'max_depth': [30]}

# Grid search
rf_cv = GridSearchCV(rf, rf_param_grid, cv=7)
rf_cv.fit(X_train, y_train)

# Record the ending time
end_time = time.time()

# Calculate the elapsed time
elapsed_time = end_time - start_time

# Print the results
print("Best Score:", rf_cv.best_score_)
print("Best Parameters:", rf_cv.best_params_)
print("Elapsed Time:", elapsed_time, "seconds")

I got good results for each class here over 98% for each:

# Make predictions on the training data
y_train_pred = rf_cv.predict(X_train)

# Compute accuracy
accuracy = accuracy_score(y_train, y_train_pred)

# Compute precision, recall, and F1-score for each class
precision = precision_score(y_train, y_train_pred, average=None)
recall = recall_score(y_train, y_train_pred, average=None)
f1 = f1_score(y_train, y_train_pred, average=None)

# Compute macro-averaged precision, recall, and F1-score
macro_precision = precision_score(y_train, y_train_pred, average='macro')
macro_recall = recall_score(y_train, y_train_pred, average='macro')
macro_f1 = f1_score(y_train, y_train_pred, average='macro')

# Print the evaluation metrics
print("Accuracy:", accuracy)
print("Precision (Class 0, 1, 2):", precision)
print("Recall (Class 0, 1, 2):", recall)
print("F1-score (Class 0, 1, 2):", f1)
print("Macro-averaged Precision:", macro_precision)
print("Macro-averaged Recall:", macro_recall)
print("Macro-averaged F1-score:", macro_f1)

The confusion matrix where it doesn't show all the labels except class 0

# Generate the confusion matrix
conf_matrix = confusion_matrix(y_train, y_train_pred)

# Define class labels
class_labels = ['Class 0', 'Class 1', 'Class 2']

# Visualize the confusion matrix with class labels
plt.figure(figsize=(8, 6))
sns.heatmap(conf_matrix, annot=True, fmt="d", cmap="Blues", xticklabels=class_labels, yticklabels=class_labels)
plt.xlabel('Predicted Label')
plt.ylabel('True Label')
plt.title('Confusion Matrix')
plt.show()

It seems like the trouble is with matplotlib/seaborn. (I can't reproduce it; maybe you need to give us a reproducible example with the exact code you wrote and your dataset.)

Instead of using the plot, you can display/print the confusion matrix as a DataFrame.

import pandas as pd
from sklearn.metrics import confusion_matrix

def get_confusion_matrix_df(classifier, X, y):
    """Return the confusion matrix as a DataFrame."""
    labels = classifier.classes_
    columns_labels = pd.MultiIndex.from_product([["Predicted"], labels])
    index_labels = pd.MultiIndex.from_product([["Actual"], labels])
    prediction = classifier.predict(X)
    matrix = confusion_matrix(y, prediction, labels=labels)
    return pd.DataFrame(matrix, columns=columns_labels, index=index_labels)

get_confusion_matrix_df(rf_cv, X_train, y_train)

Example:

from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split 
from sklearn.ensemble import RandomForestClassifier

X, y = load_iris(return_X_y=True, as_frame=True)

X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=42)

model = RandomForestClassifier(random_state=42)
model.fit(X_train, y_train)

get_confusion_matrix_df(model, X_test, y_test)

Result: