Unlock hundreds more features
Save your Quiz to the Dashboard
View and Export Results
Use AI to Create Quizzes and Analyse Results

OR
Sign inSign in with Facebook
Sign inSign in with Google
Quizzes > Computers & IT

Machine Learning Assessment Test: Check Your AI Basics

Quick, free ML knowledge test to gauge your skills. Instant results.

Editorial: Review CompletedCreated By: Evin EliasUpdated Aug 28, 2025
Difficulty: Moderate
Questions: 20
Learning OutcomesStudy Material
Colorful paper art depicting elements related to AI and Machine Learning for a knowledge quiz

This machine learning assessment quiz helps you check core concepts and find gaps to study next. After you finish, broaden your basics with the ai knowledge quiz, focus on vision skills in the computer vision quiz, or plan adoption with the free ai readiness quiz.

Which term describes an individual measurable property or characteristic used as input in a machine learning model?
Label
Parameter
Feature
Loss function
Features are individual input properties used by algorithms to make predictions. Parameters are internal model values learned during training, and labels are the target outputs. Loss functions measure prediction error rather than represent inputs.
Which machine learning task involves predicting continuous numeric values?
Dimensionality reduction
Regression
Clustering
Classification
Regression algorithms predict continuous outcomes such as price or temperature. Classification predicts discrete categories, clustering groups data without labels, and dimensionality reduction simplifies feature space.
What are datasets with known outcomes called in supervised learning?
Labeled data
Augmented data
Unlabeled data
Synthetic data
Labeled data includes both input features and their corresponding true output values needed for supervised learning. Unlabeled, synthetic, or augmented data do not refer specifically to original datasets with known targets.
What is the purpose of splitting data into training and test sets?
To perform hyperparameter tuning
To increase training speed
To improve feature selection
To evaluate model performance on unseen data
Separating data into training and test sets allows evaluation of how well a model generalizes to new, unseen examples. This helps detect overfitting. Training and test sets serve different roles in the model development pipeline.
Which of the following is an example of an unsupervised learning algorithm?
K-Means Clustering
Logistic Regression
Linear Regression
Decision Tree Classification
K-Means Clustering groups data points based on similarity without using labeled outcomes. Linear and logistic regression, as well as decision tree classification, require known target labels for training.
What is normalization in data preprocessing?
Scaling data to a 0 - 1 range
Encoding categorical variables
Transforming data to have zero mean and unit variance
Removing duplicates from data
Normalization typically refers to min - max scaling, which rescales features to a defined range such as 0 - 1. Standardization, by contrast, centers data to zero mean and unit variance. Encoding handles categorical values separately.
Which algorithm is commonly used for text classification by modeling term probabilities?
K-Means
Principal Component Analysis
Linear Regression
Naive Bayes
Naive Bayes calculates the probability of terms given a class label to classify text documents. K-Means, linear regression, and PCA are not typically used for probabilistic text classification.
Which metric is appropriate for evaluating the accuracy of a regression model?
Mean Squared Error
Accuracy
F1 Score
ROC AUC
Mean Squared Error measures the average squared difference between predicted and actual continuous values in regression tasks. Accuracy, ROC AUC, and F1 Score apply to classification problems.
What does k-fold cross-validation aim to estimate?
Training time
Number of features
Model performance on unseen data
Hyperparameter count
k-fold cross-validation splits data into k subsets to repeatedly train and validate a model, providing a robust estimate of its generalization performance. It does not directly measure training time or feature count.
In the k-nearest neighbors algorithm, what does 'k' represent?
The number of features selected
The depth of the decision tree
The number of nearest neighbors considered
The learning rate
In k-NN, 'k' specifies how many nearest neighbors are used to determine the class or value for a new data point. It does not relate to feature selection, learning rate, or tree depth.
What problem arises when a model performs well on training data but poorly on test data?
Underfitting
Data leakage
Feature scaling error
Overfitting
Overfitting occurs when a model captures noise in the training set and fails to generalize to new data. Underfitting is the opposite issue, data leakage refers to leaking test information, and scaling errors are preprocessing mistakes.
Which preprocessing step converts categorical text labels into binary indicator variables?
One-hot encoding
Label encoding
Feature scaling
Dimensionality reduction
One-hot encoding creates binary columns for each category level, enabling models to interpret categorical data. Label encoding assigns integer codes rather than separate binary features, and scaling or reduction do not handle categoricals.
What does the bias-variance tradeoff describe?
Balance between model simplicity and complexity
Relation between training and testing speed
Difference between classification and regression
Tradeoff between data volume and quality
The bias-variance tradeoff involves choosing a model that is neither too simple (high bias) nor too complex (high variance) to minimize overall error. It is not about data volume, speed, or task type.
Which method partitions data into a tree of clusters by iteratively merging or splitting?
Hierarchical clustering
K-Means clustering
Random Forest
Principal Component Analysis
Hierarchical clustering builds a nested tree of clusters by successively merging or splitting groups. K-Means fixes the number of clusters, PCA reduces dimensions, and Random Forest is an ensemble classifier.
What is the primary role of a validation set in model development?
To deploy the model
To normalize features
To train the final model
To tune hyperparameters
A validation set is used to evaluate different hyperparameter settings to prevent overfitting on the training data. It is not used to train the final model, deploy it, or for feature normalization.
When applying Principal Component Analysis (PCA), which aspect of the data is preserved most in the principal components?
Feature correlations
Mean
Variance
Label distribution
PCA transforms data into components that capture the maximum variance along orthogonal directions. While it does consider correlations implicitly, its primary goal is variance preservation, not mean centering or label distribution.
In ridge regression, increasing the regularization parameter λ will typically:
Remove outliers
Decrease variance and increase bias
Increase the learning rate
Increase variance and decrease bias
Ridge regression penalizes large coefficients, so a higher λ shrinks weights, reducing variance at the cost of increased bias. It does not affect learning rate or directly remove outliers.
What is the main difference between L1 and L2 regularization?
L1 minimizes squared error
L1 can yield sparse solutions by zeroing coefficients
L2 ignores large coefficients entirely
L2 sets coefficients to exactly one
L1 regularization (lasso) adds an absolute penalty that can force some coefficients to zero, producing sparse models. L2 (ridge) uses a squared penalty, which shrinks coefficients but typically does not set them exactly to zero.
Which ensemble technique focuses on reducing variance by averaging multiple models trained on bootstrapped samples?
Gradient Descent
Stacking
Boosting
Bagging
Bagging (bootstrap aggregating) trains models on different random samples of the data and averages their predictions to reduce variance. Boosting focuses on sequential error correction, stacking combines different algorithms, and gradient descent is an optimization method.
What does the F1 score represent in model evaluation?
The geometric mean of accuracy and recall
The harmonic mean of precision and recall
The ratio of true positives to false positives
The simple average of precision and specificity
The F1 score is defined as the harmonic mean of precision and recall, balancing the two metrics in imbalanced classification scenarios. It is neither a geometric mean of accuracy nor a simple ratio of true positives to false positives.
0
{"name":"Which term describes an individual measurable property or characteristic used as input in a machine learning model?", "url":"https://www.quiz-maker.com/QPREVIEW","txt":"Which term describes an individual measurable property or characteristic used as input in a machine learning model?, Which machine learning task involves predicting continuous numeric values?, What are datasets with known outcomes called in supervised learning?","img":"https://www.quiz-maker.com/3012/images/ogquiz.png"}

Learning Outcomes

  1. Analyse core concepts of AI and machine learning
  2. Evaluate common machine learning algorithms and their use cases
  3. Identify key terminologies and real-world AI applications
  4. Demonstrate understanding of data preprocessing techniques
  5. Apply critical thinking to select suitable ML models
  6. Master basic principles of model evaluation and tuning

Cheat Sheet

  1. Understand the Basics of AI and Machine Learning - Dive into how AI systems mimic human thinking and how machine learning teaches computers to learn from data patterns. These concepts power everything from smart assistants to recommendation engines, laying the groundwork for advanced AI projects.
  2. Differentiate Between Supervised and Unsupervised Learning - Supervised learning trains models on labeled datasets, guiding them to predict known outcomes, while unsupervised learning uncovers hidden patterns in unlabeled data without explicit instruction. Mastering this distinction helps you pick the right technique for classification, clustering, or anomaly detection challenges.
  3. Explore Key Machine Learning Algorithms - Get acquainted with popular algorithms like Linear Regression, Decision Trees, and Support Vector Machines, each designed to tackle different data challenges. Understanding their strengths, limitations, and real-world examples empowers you to choose the perfect tool for your project.
  4. Recognize the Importance of Data Preprocessing - Clean data leads to stronger models, so learn to normalize values, handle missing entries, and select meaningful features. Good preprocessing transforms raw chaos into structured insights, ensuring your algorithms learn the right patterns, not the noise.
  5. Understand Model Evaluation Metrics - Metrics like accuracy, precision, recall, and F1-score give you a comprehensive view of model performance, highlighting where your model shines and where it needs work. Balancing these measures is crucial for building reliable, real-world AI systems.
  6. Learn About Overfitting and Underfitting - Overfitting occurs when a model memorizes training data and fails on new examples, while underfitting means it's too simple to capture important patterns. Finding the sweet spot between these extremes is key to creating robust, generalizable models.
  7. Explore Real-World Applications of AI - Discover AI in action across healthcare diagnostics, financial forecasting, transportation logistics, and more - solving complex problems with data-driven magic. Seeing these examples firsthand sparks innovation and shows how theory turns into impactful solutions.
  8. Understand Neural Networks and Deep Learning - Neural networks mimic the brain's interconnected neurons, powering deep learning models that excel at tasks like image recognition and natural language processing. Delve into layers, activation functions, and backpropagation to unlock these advanced AI capabilities.
  9. Study the Concept of Reinforcement Learning - In reinforcement learning, agents learn optimal strategies by receiving rewards for good actions and penalties for mistakes, much like training a digital pet. This trial-and-error approach drives breakthroughs in gaming, robotics, and autonomous systems.
  10. Familiarize Yourself with Ethical Considerations in AI - Addressing bias, ensuring transparency, and protecting privacy are essential for responsible AI development. By embedding fairness and accountability into your models, you help ensure AI benefits everyone - now and in the future.
Make a Quiz (FREE) Copy & Edit This Quiz Create a Quiz with AI

Technology Quizzes

Powered by: Quiz Maker