Evaluation metrics (accuracy, F1, confusion matrix) in NLP - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Calculate accuracy from predictions and true labels

Given the true labels and predicted labels below, what is the accuracy of the model?

NLP

true_labels = [1, 0, 1, 1, 0, 0, 1]
pred_labels = [1, 0, 0, 1, 0, 1, 1]
correct = sum(t == p for t, p in zip(true_labels, pred_labels))
accuracy = correct / len(true_labels)
print(round(accuracy, 2))

A 0.85

B 0.57

C 0.71

D 0.43

💡 Hint

Count how many predictions match the true labels, then divide by total labels.

Accuracy is the number of correct predictions divided by total predictions. Here, 5 out of 7 predictions are correct, so accuracy is 5/7 ≈ 0.71.

🧠 Conceptual

intermediate

Understanding the F1 score

Which statement best describes the F1 score in classification tasks?

A It is the average of accuracy and recall.

B It measures the ratio of correct predictions to total predictions.

C It counts the number of true negatives in the confusion matrix.

D It is the harmonic mean of precision and recall, balancing both.

💡 Hint

Think about how F1 combines two important metrics to balance false positives and false negatives.

F1 score combines precision and recall using the harmonic mean, giving a single metric that balances both false positives and false negatives.

❓ Predict Output

advanced

Confusion matrix values from predictions

Given the true and predicted labels below, what is the value of True Positives (TP) in the confusion matrix?

NLP

true_labels = [0, 1, 1, 0, 1, 0, 1, 1]
pred_labels = [0, 1, 0, 0, 1, 1, 1, 0]
TP = sum(1 for t, p in zip(true_labels, pred_labels) if t == 1 and p == 1)
print(TP)

A 3

B 2

C 4

D 1

💡 Hint

Count how many times the true label and predicted label are both 1.

True Positives are cases where the true label is 1 and the prediction is also 1. Here, that happens three times.

❓ Model Choice

advanced

Choosing metric for imbalanced data

You have a dataset where 95% of samples belong to class A and 5% to class B. Which metric is best to evaluate your model's performance on class B?

A F1 score

B Loss value

C Accuracy

D Total number of predictions

💡 Hint

Think about which metric balances false positives and false negatives well for rare classes.

Accuracy can be misleading with imbalanced data. F1 score balances precision and recall, making it better for rare classes.

❓ Metrics

expert

Calculate macro F1 score from class-wise precision and recall

Given the following precision and recall for three classes, what is the macro F1 score? Class 1: precision=0.8, recall=0.6 Class 2: precision=0.7, recall=0.7 Class 3: precision=0.9, recall=0.5

A 0.72

B 0.67

C 0.70

D 0.75

💡 Hint

Calculate F1 for each class, then average them.

F1 for each class = 2 * (precision * recall) / (precision + recall). Then average all three F1 scores for macro F1.