Classification Quiz

What is the difference between classification and regression?

• Classification uses numerical targets, regression uses categories

• Classification predicts continuous variables, regression predicts labels

• Classification predicts categorical outputs, regression predicts numerical outputs

• Classification models don’t require training data

Which of the following is an example of a classification problem?

• Predicting next month’s electricity bill

• Predicting the stock market index

• Predicting whether a customer will churn

• Estimating the height of a plant

What is required for a supervised classification model to be trained?

• Labelled data

• Unlabelled data

• Time series data

• Only numerical data

Which step comes first in the classification workflow?

• Model evaluation

• Algorithm implementation

• Exploratory Data Analysis

• Prediction

Multi-class classification refers to problems where

• There are two output classes

• The target variable is numerical

• Unsupervised learning is used

• There are more than two possible output categories

Why might it be a problem if a dataset has significantly more of one category than another when training a model?

• It will take longer to train the model

• The model may become biased toward predicting the majority class

• The model cannot run on imbalanced data

• The model might underfit the data

Which metric combines both precision and recall?

• F1 Score

• Accuracy

• Confusion Matrix

• ROC AUC Score

What does the max_depth parameter control in a decision tree classifier?

• The number of features

• The number of output labels

• The maximum number of rows

• The depth of the tree, controlling complexity

When should you prefer Random Search over Grid Search for hyperparameter tuning?

• When the dataset is small

• When you have few hyperparameters

• When the parameter space is large and resources are limited

• When you only have categorical features

What is the benefit of using probability predictions over direct class labels?

• They show the confidence of the model

• They are faster to compute

• They help in data cleaning

• They reduce overfitting