Classification and Decision Boundaries Quiz

Classification problems involve predicting:

• Continuous numerical values

• Categories or discrete labels

• Missing data

• Random noise

Which of the following is a binary classification problem?

• Predicting price of a car

• Predicting if an email is spam or not

• Predicting the species of a flower

• Predicting rainfall in cm

Which of these algorithms can be used for classification?

• Linear Regression

• K-Means

• Logistic Regression

• PCA

The output variable in classification is:

• Continuous

• Discrete

• Real-valued

• None of these

The main goal of a classification algorithm is to:

• Minimize reconstruction error

• Find clusters in unlabeled data

• Learn decision boundaries that separate classes

• Maximize variance of data

A decision boundary separates:

• The training and test data

• Different features

• Different classes in the feature space

• Outliers from inliers

For a 2D binary classifier, the decision boundary is usually a:

• Point

• Line

• Circle

• Cube

In linear classification, decision boundary satisfies: f(x)=w1x1+w2x2+b=0 What does this equation represent?

• The dividing line between two classes

• The loss function

• The gradient of the model

• The cost surface

When the decision boundary is non-linear, which classifier is more suitable?

• Linear Regression

• Logistic Regression

• SVM with kernel trick

• Simple perceptron

A linear classifier uses which function form?

• f(x)=wTx+b

• f(x)=sin(x)

• f(x)=ex

• f(x)=x2