Machine Learning Quiz

Questions: 16 · 10 minutes
1. A table contains house size, neighborhood, year built, and final sale price. If the goal is to predict sale price, which column is the target label?
Final sale price
Neighborhood
House size
Year built
2. During neural-network training, the loss jumps up and down and fails to settle because parameter updates repeatedly overshoot useful values. Which adjustment is most likely to help?
Remove the validation set.
Replace all numeric features with category labels.
Reduce the learning rate.
Increase the learning rate substantially.
3. A retailer wants to identify groups of customers with similar purchasing patterns, but it has no predefined customer categories. Which method is most suitable?
Regression
Supervised classification
Reinforcement learning
Clustering
4. A team has thousands of emails already labeled “spam” or “not spam.” Which approach most directly fits the task of predicting the label for new emails?
Unsupervised clustering
Reinforcement learning
Supervised classification
Dimensionality reduction
5. A practitioner wants to compare model settings without repeatedly using the final test set. What is the main benefit of cross-validation on the training data?
It proves that the model will perform equally well in every future setting.
It permanently eliminates bias from the training data.
It converts an unsupervised problem into a supervised one.
It estimates performance across multiple training-validation splits.
6. Which statement best defines supervised learning?
A model learns from examples that include known target labels or values.
A model searches unlabeled data for naturally occurring groups.
An agent learns only by receiving rewards from an environment.
A model generates labels for data without using any training objective.
7. In a rare-disease screening model, both missed cases and excessive false alarms matter. Which metric combines precision and recall into one score?
Mean squared error
F1 score
R-squared
Silhouette score
8. Before splitting a dataset, an analyst calculates normalization values from the entire dataset and then uses them during training. Why can this be a problem?
Normalization can be used only with clustering algorithms.
The procedure necessarily removes the target variable.
The procedure makes the training set too small to fit a model.
Information from the future test set can leak into model development.
9. A model must predict the expected electricity consumption of a building in kilowatt-hours. What kind of supervised task is this?
Classification
Clustering
Regression
Association-rule learning
10. A random forest produces a prediction by combining the outputs of many decision trees. What broad machine learning idea does this illustrate?
Feature scaling
Dimensionality reduction
Reward shaping
Ensemble learning
11. A dataset has hundreds of correlated numeric features, and an analyst wants a smaller set of components that preserves as much variation as possible. Which technique is designed for this?
K-nearest neighbors
Principal component analysis
Logistic regression
Q-learning
12. How does deep learning relate to machine learning?
It is a separate field that does not use data-driven optimization.
It is a machine learning approach based on neural networks with multiple representation layers.
It refers only to decision trees trained on very large tables.
It is another name for any unsupervised clustering method.
13. A company has a small set of labeled images and a much larger set of unlabeled images. It wants to use both sets while learning to classify new images. Which main approach best matches this plan?
Ordinary supervised learning using only labeled examples
Semi-supervised learning
Reinforcement learning
Pure clustering with no use of labels
14. Which situation is the clearest example of reinforcement learning?
Training a game-playing agent through rewards for successful actions
Grouping news articles by similarity without labels
Predicting whether a transaction is fraudulent from labeled records
Estimating apartment prices from past sales
15. A model scores 99% on its training data but only 68% on new test data drawn from the same setting. What is the most likely problem?
Underfitting
Clustering instability
Overfitting
A reinforcement reward delay
16. Why is a test set normally kept separate from the data used to fit a model?
To estimate how well the fitted model generalizes to unseen data
To increase the number of features available during training
To guarantee that every class has the same number of examples
To ensure the model achieves perfect training accuracy
Popular tests
Narcissistic Personality Inventory (NPI)
This self-report measure is used to assess narcissism as a personality trai…
Start Test
Yale-Brown Obsessive Compulsive Scale (Y-BOCS)
This measure is used to rapidly quantify the current severity of obsessive…
Start Test
CRAFFT Screening Test (CRAFFT 2.1)
This brief screening measure is designed to identify potential alcohol and…
Start Test
Patient Health Questionnaire-9 (PHQ-9)
This measure is commonly used to quickly screen for the presence and severi…
Start Test
Maslach Burnout Inventory (MBI)
This self-report measure is used to assess occupational burnout symptoms in…
Start Test
Adolescent Anxiety Questionnaire
This measure is designed to support a brief appraisal of anxiety symptoms a…
Start Test
Emotional Creativity Inventory (ECI)
This self-report measure assesses individual differences in the originality…
Start Test
Horne–Ostberg Morningness–Eveningness Questionnaire (MEQ)
Circadian preferences influence typical patterns of alertness and sleep tim…
Start Test
Ambivalent Sexism Inventory (ASI)
This measure is designed to assess attitudes toward women, including both o…
Start Test
Internalized Misogyny Scale (IMS)
This measure is designed to assess internalized negative beliefs and stereo…
Start Test
Perceived Stress Scale (PSS-10)
This self-report measure assesses the degree to which individuals appraise…
Start Test
Impulsive Behavior Scale (SUPPS-P)
Impulsivity is a multidimensional construct that is often assessed with bri…
Start Test
Clinical Institute Withdrawal Assessment for Alcohol, Revised (CIWA-Ar)
This rating scale is used to rapidly assess the severity of alcohol withdra…
Start Test
Positive and Negative Affect Schedule (PANAS)
This measure provides a brief self-report assessment of current or typical…
Start Test
Light Triad Scale (LTS)
This self-report measure assesses prosocial personality tendencies and orie…
Start Test
Suicidal Ideation Scale
In clinical settings, the Suicidal Ideation Scale is used to structure an i…
Start Test
Body Dysmorphic Disorder Scale (BDD-D)
This brief self-report measure is designed to screen for and quantify distr…
Start Test
Beck Anxiety Inventory (BAI)
This measure is a brief self-report inventory used to screen for anxiety sy…
Start Test
Differential Test of Perfectionism
This instrument is used to screen for perfectionism-related attitudes and t…
Start Test
Locus of Control Scale
This measure assesses generalized expectancies regarding the degree to whic…
Start Test
New Apathy Scale
This brief self-report measure is used to screen for apathy-related symptom…
Start Test
Perth Alexithymia Questionnaire (PAQ)
This measure assesses individual differences in alexithymia, including diff…
Start Test
Social Intelligence Scale
This brief self-report measure is designed to support rapid screening of in…
Start Test
Fear Test
This measure is designed to evaluate individual differences in fear-related…
Start Test
Neuroticism Level Scale
The measure is intended for brief screening of an individual’s propensity t…
Start Test
Aggressiveness Indicators Screening Questionnaire
This screening tool is designed to quickly identify behavioral indicators a…
Start Test