Multiple Choice Quiz on Classification Models

A confusion matrix is primarily used to understand where a classification model gets confused.

• Trained

• Optimized

• Confused

• Deployed

Key performance measures that can be calculated from a confusion matrix include accuracy, precision, and recall.

• Speed, latency, and throughput

• Cost, efficiency, and scalability

• Accuracy, precision, and recall

• Stability, robustness, and fairness

A true negative (TN) in a confusion matrix represents a case where the model correctly predicted a negative outcome.

• Incorrectly predicted a positive outcome

• Incorrectly predicted a negative outcome

• Correctly predicted a positive outcome

• Correctly predicted a negative outcome

A Type I error is synonymous with a false positive (FP).

• True positive

• True negative

• False negative

• Type I error

A false negative (FN) occurs when the model incorrectly predicts a negative outcome, but the actual outcome was positive.

• Correctly predicts a positive outcome

• Correctly predicts a negative outcome

• Incorrectly predicts a positive outcome

• Incorrectly predicts a negative outcome, but the actual outcome was positive

Accuracy, as an evaluation metric, calculates how often the model's predictions are correct overall.

• The proportion of correctly predicted positive instances

• The proportion of actual positive instances correctly identified

• How often the model's predictions are correct overall

• The balance between positive and negative predictions

Accuracy can be a misleading metric when dealing with imbalanced datasets.

• Large datasets

• Datasets with many features

• Imbalanced datasets

• Datasets with normally distributed classes

Classification models are primarily evaluated using accuracy.

• Regression models

• Clustering algorithms

• Dimensionality reduction techniques

• Classification models

Accuracy is not an appropriate evaluation metric for regression tasks.

• Binary classification

• Multi-class classification

• Regression

• Image recognition

For binary classification, accuracy is calculated as (TP + TN) / (TP + FP + TN + FN).

• TP / (TP + FP)

• TP / (TP + FN)

• (TP + FP) / (TP + FP + TN + FN)

• (TP + TN) / (TP + FP + TN + FN)