When is accuracy misleading?

When is accuracy misleading?

• A. When datasets are imbalanced
• B. When the dataset is small
• C. When there are many classes
• D. When there are no misclassifications

Answer: (A)

Accuracy can be misleading when the data are imbalanced. If one class dominates the dataset, a model can achieve high accuracy simply by predicting that majority class every time, while completely failing on the minority class. For example, with 100 samples where 95 are of one class and 5 are another, a model that always predicts the majority class would be correct 95 times, giving 95% accuracy, but it would miss all of the minority cases. That apparent high performance hides the model’s poor ability to detect the less frequent class. In imbalanced situations, other metrics like precision, recall, F1 score, or AUC-ROC provide a clearer view of how well the model performs across all classes.

Accuracy can be misleading when the data are imbalanced. If one class dominates the dataset, a model can achieve high accuracy simply by predicting that majority class every time, while completely failing on the minority class. For example, with 100 samples where 95 are of one class and 5 are another, a model that always predicts the majority class would be correct 95 times, giving 95% accuracy, but it would miss all of the minority cases. That apparent high performance hides the model’s poor ability to detect the less frequent class. In imbalanced situations, other metrics like precision, recall, F1 score, or AUC-ROC provide a clearer view of how well the model performs across all classes.