INTERVIEW QUESTIONS AND ANSWERS OF MACHINE LEARNING
1. Differentiate between Supervised, Unsupervised, and Reinforcement
Learning.
Answer:
Supervised Learning: Uses labeled data (e.g., regression, classification).
Unsupervised Learning: Uses unlabeled data to identify patterns (e.g., clustering).
Reinforcement Learning: An agent learns by interacting with an environment and
receiving rewards.
2. What are the types of Machine Learning algorithms?
Answer:
1. Classification (e.g., logistic regression).
2. Regression (e.g., linear regression).
3. Clustering (e.g., K-means).
4. Dimensionality Reduction (e.g., PCA).
5. Anomaly Detection.
3. Explain Overfitting and Underfitting.
Answer:
Overfitting: The model performs well on training data but poorly on unseen data.
Underfitting: The model performs poorly on both training and unseen data.
4. What is Gradient Descent?
Answer:
Gradient Descent is an optimization algorithm that minimizes a cost function by updating
model parameters iteratively in the direction of the negative gradient.
5. What is a Support Vector Machine (SVM)?
Answer:
SVM is a supervised learning algorithm that finds the hyperplane maximizing the margin
between classes in a dataset.
6. Explain the Difference Between Linear and Logistic Regression.
Answer:
Linear Regression: Predicts a continuous value using a linear relationship between input
features and the target.
Logistic Regression: Predicts probabilities for classification tasks and outputs a value
between 0 and 1 using the sigmoid function.
7. What is Bayesian Learning?
Answer:
Bayesian learning is a probabilistic approach in machine learning where Bayes' theorem is
used to update the probability of a hypothesis as new data is observed.
8. How does Bayes' Theorem work in Bayesian Learning?
Answer:
Bayes' theorem calculates the posterior probability of a hypothesis
P(A|B) = P(B|A)P(A) / P(B)
where,
P(A) and P(B) are the probabilities of events A and B
P(A|B) is the probability of event A when event B happens
P(B|A) is the probability of event B when A happens
9. What are the common types of SVM Kernels?
Answer:
1. Linear Kernel: Suitable for linearly separable data.
2. Polynomial Kernel: Captures nonlinear relationships with polynomial terms.
3. RBF (Gaussian) Kernel: Handles complex nonlinear separations.
4. Sigmoid Kernel: Similar to neural networks.
10. How does a Decision Tree split the data?
Answer:
A decision tree splits the data using measures like:
Gini Index: Measures impurity in splits.
Entropy (Information Gain): Measures information gain for a split.
Variance Reduction: Used for regression tasks.
10. What is Overfitting in Decision Trees, and how can it be avoided?
Answer:
Overfitting occurs when a decision tree becomes too complex and performs well on training
data but poorly on unseen data.
Avoidance Techniques:
1. Pruning (pre-pruning or post-pruning).
2. Limiting tree depth.
3. Setting a minimum sample size for splits.
11. What are some examples of Instance-Based Learning algorithms?
Answer:
1. k-Nearest Neighbors (k-NN).
2. Locally Weighted Regression.
3. Case-Based Reasoning.
4. Radial Basis Function Networks.
12. What is Backpropagation in ANN?
Answer:
Backpropagation is a learning algorithm that computes the gradient of the loss function with
respect to weights, propagating errors backward through the network to update weights using
gradient descent.
13. What are Convolutional Neural Networks (CNNs) used for?
Answer:
CNNs are used for image and video processing tasks, including image classification, object
detection, and segmentation, by learning spatial hierarchies of features.
14. Explain the key components of a Reinforcement Learning system.
Answer:
1. Agent: The learner or decision-maker.
2. Environment: The system with which the agent interacts.
3. State (S): A representation of the environment's current condition.
4. Action (A): Choices made by the agent.
5. Reward (R): Feedback signal to evaluate actions.
6. Policy (π): The strategy the agent follows to make decisions.
7. Value Function (V): Estimates future rewards from a state.
15. What is a Markov Decision Process (MDP)?
Answer:
MDP provides a mathematical framework for modeling RL problems with the following
components:
1. States (S).
2. Actions (A).
3. Transition probabilities (P).
4. Rewards (R).
5. Discount factor (γ).
