Machine Learning Cheat Sheet 2026

The 30 highest-yield Machine Learning facts, distilled from real exam questions. Print it, save it as a PDF, or study it here — free, no sign-up.

1. In Machine Learning practice, what is the best approach to quality improvement in computer vision? → Use data-driven methods with measurable outcomes
2. Which of the following is NOT a hyperparameter of a Random Forest model? → Learned feature weights of the trees
3. What does the term 'epoch' refer to in neural network training? → One complete pass through the entire training dataset
4. A classifier achieves 98% accuracy on a dataset where 98% of samples belong to class A. What problem does this illustrate? → The accuracy paradox with imbalanced data
5. Which metric is commonly used to evaluate machine translation quality by comparing n-gram overlap with reference translations? → BLEU score
6. Which of the following clustering algorithms is specifically designed to handle arbitrarily shaped clusters? → DBSCAN
7. What does the R² (coefficient of determination) value of 0 indicate? → The model performs as well as a constant mean predictor
8. What is model quantization in the context of deploying deep learning models? → Reducing the numerical precision of weights (e.g., float32 → int8) to speed up inference
9. Which technique randomly sets neuron outputs to zero during training to prevent overfitting? → Dropout
10. In a confusion matrix for binary classification, what does a 'False Positive' represent? → A negative instance incorrectly classified as positive
11. Which clustering approach would be most appropriate for customer segmentation when the true number of segments is unknown and data has irregular shapes? → DBSCAN with tuned epsilon and minPts
12. What does 'batch normalization' do in a neural network? → Normalizes the inputs of each layer to have zero mean and unit variance during training
13. Which of the following is an example of a 'date-derived' feature useful for forecasting retail sales? → Day of the week
14. What is the primary advantage of using an ensemble method like Gradient Boosting over a single decision tree? → Higher predictive accuracy by combining many weak learners sequentially
15. What is the most important professional competency for Machine Learning certification in computer vision? → Deep knowledge combined with practical application skills
16. In spectral clustering, which step transforms the data before applying K-Means? → Mapping data into eigenvector space of the graph Laplacian
17. What is the primary consideration when implementing changes to neural networks? → Impact assessment and change management
18. In a Random Forest classifier, how does each tree differ from the others? → Each tree is trained on a bootstrap sample with a random subset of features at each split
19. What is the primary purpose of the softmax function in the output layer of a classifier? → To convert raw logits into a probability distribution summing to 1
20. Which of the following is a feature extraction example? → All of the above
21. Which cross-validation approach is most appropriate for time-series data? → Walk-forward (time-series) cross-validation
22. In-text mining, which of the following approaches can be used for normalization? → Both B and C
23. Which technique is used to select the optimal regularization parameter in Ridge or Lasso regression? → Cross-validation
24. Which loss function is typically used for binary classification tasks in deep learning? → Binary cross-entropy
25. What is the primary advantage of using 'embeddings' over one-hot encoding for high-cardinality categorical variables? → Embeddings produce dense, low-dimensional representations that capture semantic similarity
26. What is 'spectral clustering' and what advantage does it have? → It uses eigenvectors of a similarity graph Laplacian to cluster data of non-convex shapes
27. Which phenomenon occurs when a classifier performs well on training data but fails to generalize to unseen examples? → Overfitting
28. Which of the following is an example of a generative classification model? → Naive Bayes
29. Which approach best demonstrates mastery of classification in Machine Learning practice? → Applying principles to novel situations with sound judgment
30. What is the purpose of 'feature crosses' in models like logistic regression? → To allow linear models to learn non-linear decision boundaries by combining features