Here are 25 important machine learning interview questions along with short answers:
Question | Answer |
---|---|
What is machine learning? | AI training for predictions based on data. |
Supervised vs. unsupervised learning? | Labeled vs. unlabeled data for model training. |
Classification vs. regression? | Predicting categories vs. continuous values. |
What is overfitting? | Model fits data closely, fails on new data; prevent. |
What is cross-validation? | Testing model by splitting data multiple times. |
What is a decision tree? | Tree-like model for decisions using features. |
What is a neural network? | Brain-inspired model with layers for data processing. |
What is regularization? | Prevents overfitting by adding penalty to loss function. |
What is gradient descent? | Optimizes models by adjusting parameters for minimum loss. |
What is deep learning? | Trains neural networks with layers for complex patterns. |
What is a support vector machine (SVM)? | Used for classification and regression. |
Parametric vs. non-parametric model? | Fixed vs. infinite parameters learned from data. |
What is the curse of dimensionality? | High-dimensional data sparsity makes pattern finding hard. |
What is principal component analysis (PCA)? | Reduces dimensions by finding feature combinations. |
What is k-fold cross-validation? | Splits data into k subsets for testing and training. |
Precision vs. recall? | Measures accuracy of positive predictions vs. true positives. |
What is the F1 score? | Balances precision and recall for model evaluation. |
Bias-variance tradeoff? | Balances model complexity for better generalization. |
What is transfer learning? | Uses pre-trained models to save time and improve performance. |
What is ensemble learning? | Combines multiple models for better predictions. |
What is deep reinforcement learning? | Combines deep learning with reinforcement for decision making. |
Batch vs. stochastic gradient descent? | Updating model with entire data vs. one data point. |
L1 vs. L2 regularization? | Absolute value vs. square value penalties for model parameters. |
Generative vs. discriminative model? | Joint vs. conditional probability distributions for tasks. |
CNN vs. RNN? | Image recognition vs. sequence data processing in neural networks. |
Please write comments below if you find anything incorrect, or you want to share more information about the topic discussed above. A gentle request to share this topic on your social media profile.