Which regularisation methods are affected by features on larger scales receiving disproportionately small penalties?

Decision Trees, Random Forests, XGBoost

What is a characteristic of tree-based models like Decision Trees, Random Forests, and XGBoost regarding feature scaling?

They are generally scale-invariant.

They are highly sensitive to feature scaling.

They require normalization for optimal performance.

They split based on distances, not thresholds.

What is the primary purpose of feature scaling as shown in the visualization?

To equalize the contribution of all features.

To increase the dominance of certain features.

To improve the accuracy of predictions.

After scaling, what is the relative contribution of each feature to distance/cost computation?

97% for Income, 16% for Cholesterol, 6% for Blood Pressure, 1.6% for BMI, and 0.03% for Age.

50% for Income and 10% for other features.

30% for Cholesterol and 70% for Income.

Why does Income dominate the feature contribution before scaling?

Because it has a larger range of values compared to other features.

Because it is the most important feature in the dataset.

Because it is scaled to have a higher weight.

Because it is unrelated to other features.

What is the primary purpose of feature scaling in machine learning?

To transform all features to a common, comparable numerical range

To increase the number of features in a dataset

To eliminate irrelevant features from the dataset

How does feature scaling contribute to faster convergence during gradient descent?

By creating a well-conditioned loss landscape

By increasing the number of iterations required

By reducing the number of features in the dataset

By eliminating irrelevant features

Which of the following is a benefit of stable regularization achieved through feature scaling?

Lasso and Ridge penalties are applied consistently across all features regardless of original magnitude

It eliminates the need for regularization techniques

It reduces the number of features in the dataset

It increases the complexity of the model

What does fair distance computation ensure in feature scaling?

All features contribute equally to k-NN, k-Means, and SVM distance calculations

Only the largest features are considered in distance calculations

Distance calculations are eliminated from the model

Features are ranked based on their importance

How does feature scaling improve interpretability in linear models?

By making coefficient magnitudes directly comparable across scaled features

By reducing the number of features in the model

By eliminating irrelevant features from the dataset

By increasing the complexity of the model

What is the formula for Min-Max Scaling?

x' = (x - x_min) / (x_max - x_min)

x' = (x_max - x_min) / (x - x_min)

x' = (x - x_max) / (x_min - x_max)

x' = (x_min - x) / (x_max - x_min)

What property of Min-Max Scaling ensures that the original distribution shape is preserved?

It applies exponential scaling.

It normalizes data using the mean.

What is one of the conditions under which Min-Max scaling should be used?

When values need to be strictly bounded to [0, 1].

When the data contains extreme outliers.

When the algorithm is insensitive to absolute value ranges.

When zero-point relationships are irrelevant.

Which of the following is a limitation of Min-Max scaling?

It assumes a fixed range for test data points.

It guarantees a normal distribution of data.

It is insensitive to new data changes.

It eliminates outlier sensitivity.

Why might Min-Max scaling fail when dealing with extreme outliers?

It destroys variance information by compressing values into a tiny sub-range.

It compresses all values into a large range.

It produces values outside the range [0, 1].

It guarantees Gaussian distribution.

What is a key characteristic of Min-Max scaling when preserving zero-point relationships?

It ensures that x=0 remains meaningful.

It ignores zero-point relationships.

It compresses zero-point relationships into a sub-range.

It assumes zero-point relationships are irrelevant.

Which of the following algorithms is sensitive to absolute value ranges, making Min-Max scaling useful?

Naive Bayes and decision trees.

Support vector machines and random forests.

Gradient boosting and ensemble methods.

What does the variable "x" represent in the Z-score standardization formula?

Original feature value for a given sample

Mean of the feature across all training samples

Standard deviation of the feature across all training samples

Standardized value (number of standard deviations from the mean)

What does the variable "σ" (sigma) represent in the Z-score standardization formula?

Standard deviation of the feature across all training samples

Mean of the feature across all training samples

Original feature value for a given sample

Standardized value (number of standard deviations from the mean)

Which of the following is a key property of Z-score standardization?

Outliers remain in the data but are expressed as large |z| values

Output has mean = 1 and standard deviation = 0

Output range is bounded between [0, 1]

Output has mean = 0 and standard deviation = 2

What does the variable "μ" (mu) represent in the Z-score standardization formula?

Mean of the feature across all training samples

Standard deviation of the feature across all training samples

Original feature value for a given sample

Standardized value (number of standard deviations from the mean)

What does standardization map your data onto in the context of normal distribution?

A standard normal distribution N(0,1)

Which of the following steps is required to calculate the z-score for a given age in the dataset?

Subtract the mean from the age and divide by the standard deviation.

Add the mean to the age and divide by the standard deviation.

Multiply the mean by the age and divide by the standard deviation.

Subtract the standard deviation from the age and divide by the mean.

What is the output range of the Min-Max scaling method?

Approx. -3 to +3 for normal data

What is the rule of thumb for choosing between Min-Max and Z-Score scaling methods?

Default to Z-Score unless [0, 1] inputs are explicitly required

Use Z-Score for neural networks

What is the primary issue when training and evaluating a model on the same data without splitting?

The model memorizes the data and fails to generalize.

The model learns patterns from the data.

The test set is hidden during training.

The performance gap reveals overfitting.

Which of the following is a benefit of using a test set during model evaluation?

The evaluation on the test set gives honest performance.

The model sees all data during training.

Reported accuracy is artificially optimistic.

The model memorizes the test set.

What does the performance gap between the training set and test set indicate?

The model has memorized the data.

The model has learned patterns from the data.

Artificially optimistic accuracy.

Why is splitting data into training and test sets important for evaluating a model?

It ensures the test set is hidden during training.

It allows the model to memorize the data.

It artificially inflates reported accuracy.

It prevents the model from learning patterns.

What happens when a model is evaluated on the same data it was trained on?

The reported accuracy/error is artificially optimistic.

The model learns patterns from unseen data.

The test set is hidden during training.

Honest performance metrics are obtained.

Day 2 Agenda Quiz

Authored by andres quintero

Mathematics

University

CCSS covered

Used 2+ times

AI Actions

Add similar questions

Adjust reading levels

Convert to real-world scenario

Translate activity

More...

Content View

Student View

76 questions

Show all answers

MULTIPLE CHOICE QUESTION

30 sec • 1 pt

What is the primary purpose of feature scaling in data preprocessing?

To increase the size of the dataset

To address unequal feature ranges and improve model performance

To remove missing values from the dataset

To reduce the number of features in the dataset

Which scaling method uses the formula to transform data into a range between 0 and 1?

Z-Score Standardization

Min-Max Scaling

Log Transformation

Feature Selection

What is a key advantage of Z-Score Standardization compared to Min-Max Scaling?

It transforms data into a range between 0 and 1

It uses the mean and standard deviation to standardize data

It removes outliers from the dataset

It is faster to compute than Min-Max Scaling

Why is train/test splitting important in machine learning?

To increase the size of the training dataset

To prevent data leakage and evaluate model performance

To remove irrelevant features from the dataset

To ensure all data is used for training

MULTIPLE CHOICE QUESTION

30 sec • 1 pt

Why is feature scaling important in machine learning algorithms, as highlighted in the document?

To reduce the size of the dataset

To ensure features with larger ranges do not dominate smaller ones

To eliminate irrelevant features

To increase the number of features

MULTIPLE CHOICE QUESTION

30 sec • 1 pt

Which type of algorithms are dominated by the highest-range feature, where a small change in one feature can massively outweigh changes in another?

Distance-Based Algorithms

Gradient Descent

Regularisation Methods

Tree-Based Models

MULTIPLE CHOICE QUESTION

30 sec • 1 pt

What happens to the loss landscape in Gradient Descent when feature scaling is not applied?

It becomes elongated, causing slow or unstable convergence.

It becomes flat, leading to faster convergence.

It remains unaffected by scaling.

It becomes circular, ensuring stable convergence.

Access all questions and much more by creating a free account

Create resources

Host any resource

Get auto-graded reports

Continue with Google

Continue with Email

Continue with Classlink

Continue with Clever

or continue with

Microsoft

Apple

Others

Already have an account?

Popular Resources on Wayground

25 questions

The Ultimate College Knowledge Quiz

Quiz

•

8th Grade

20 questions

Math Review

Quiz

•

3rd Grade

15 questions

Fast food

Quiz

•

7th Grade

20 questions

Math Review

Quiz

•

6th Grade

20 questions

Context Clues

Quiz

•

6th Grade

20 questions

Inferences

Quiz

•

4th Grade

19 questions

Classifying Quadrilaterals

Quiz

•

3rd Grade

20 questions

Figurative Language Review

Quiz

•

6th Grade

Discover more resources for Mathematics

24 questions

5th Grade Math EOG Review

Quiz

•

KG - University

31 questions

Math 1 EOC review

Quiz

•

KG - University