✅ Machine Learning Interview Questions & Answers 🎯

1. What is the difference between supervised and unsupervised learning
Answer:
Supervised learning uses labeled data to learn a mapping from inputs to outputs (e.g., predicting house prices). Unsupervised learning finds hidden patterns or groupings in unlabeled data (e.g., customer segmentation using K-Means).

2. How do you handle missing values during feature engineering
Answer:
Common strategies include:
– Imputation: Fill missing values with mean, median, or mode
– Deletion: Remove rows or columns with excessive missing data
– Model-based: Use predictive models to estimate missing values

3. What is the bias-variance tradeoff
Answer:
Bias refers to error due to overly simplistic assumptions; variance refers to error due to model sensitivity to small fluctuations in training data. A good model balances both to avoid underfitting (high bias) and overfitting (high variance).

4. Explain how Random Forest reduces overfitting
Answer:
Random Forest uses bagging (bootstrap aggregation) and builds multiple decision trees on random subsets of data and features. It averages their predictions, reducing variance and improving generalization.

5. What is the role of cross-validation in model selection
Answer:
Cross-validation (e.g., k-fold) splits data into multiple training/testing sets to evaluate model performance more reliably. It helps prevent overfitting and ensures the model generalizes well to unseen data.

6. How does XGBoost differ from traditional boosting methods
Answer:
XGBoost uses gradient boosting with regularization (L1 and L2), tree pruning, and parallel processing. It’s faster and more accurate than traditional boosting algorithms like AdaBoost.

7. What is the difference between L1 and L2 regularization
Answer:
– L1 (Lasso): Adds absolute value of weights to loss function, promoting sparsity
– L2 (Ridge): Adds squared value of weights, penalizing large weights and improving stability

8. How would you deploy a trained ML model
Answer:
– Serialize the model using pickle or joblib
– Create a REST API using Flask or FastAPI
– Monitor performance using metrics like latency, accuracy drift, and feedback loops

9. What is the difference between precision and recall
Answer:
– Precision: True Positives / (True Positives + False Positives)
– Recall: True Positives / (True Positives + False Negatives)
Precision focuses on correctness of positive predictions; recall focuses on capturing all actual positives.

10. What is the Q-value in reinforcement learning
Answer:
Q-value represents the expected cumulative reward of taking an action in a given state and following a policy thereafter. It’s central to Q-learning algorithms.

❤️ Tap for more