Basics of Machine Learning

1. What is machine learning?
   • Answer: Machine learning is a branch of artificial intelligence (AI) that enables systems to learn from data and improve performance over time without explicit programming.
2. Differentiate between supervised and unsupervised learning.
   • Answer:
     • Supervised Learning: Involves training a model on labeled data to make predictions or classify new data.
     • Unsupervised Learning: Involves training on unlabeled data to discover patterns, clusters, or relationships.
3. Explain the bias-variance trade-off in machine learning.
   • Answer: The bias-variance trade-off balances model complexity (variance) and generalization ability (bias). High bias leads to underfitting, and high variance leads to overfitting.

Machine Learning Algorithms

1. Name some popular machine learning algorithms and their applications.
   • Answer: Algorithms include:
     • Linear Regression: Predicting continuous outcomes.
     • Decision Trees: Classification and regression tasks.
     • Random Forest: Ensemble learning for classification and regression.
     • K-Nearest Neighbors (KNN): Classification based on nearest data points.
     • Support Vector Machines (SVM): Classification and regression using hyperplanes.
2. Explain the difference between classification and regression.
   • Answer:
     • Classification: Predicts categorical outcomes or class labels.
     • Regression: Predicts continuous numerical values.
3. What is overfitting in machine learning, and how can it be prevented?
   • Answer: Overfitting occurs when a model learns noise or irrelevant patterns in the training data, leading to poor generalization on new data. Prevention methods include cross-validation, regularization, and reducing model complexity.

Evaluation Metrics in Machine Learning

1. Name some evaluation metrics for classification models.
   • Answer: Metrics include accuracy, precision, recall, F1-score, ROC-AUC (Receiver Operating Characteristic - Area Under Curve), and confusion matrix.
2. How is accuracy calculated for a machine learning model?
   • Answer: Accuracy measures the proportion of correctly predicted instances among all predictions, calculated as (TP + TN) / (TP + TN + FP + FN), where TP = True Positive, TN = True Negative, FP = False Positive, FN = False Negative.
3. What is precision and recall? How are they related?
   • Answer:
     • Precision: Measures the proportion of correctly predicted positive instances among all predicted positives, calculated as TP / (TP + FP).
     • Recall: Measures the proportion of correctly predicted positive instances among all actual positives, calculated as TP / (TP + FN).
     • Precision and recall are inversely related; improving one often reduces the other.

Feature Engineering and Data Preprocessing

1. Why is feature scaling important in machine learning?
   • Answer: Feature scaling standardizes the range of features, preventing some features from dominating others, and improves the performance of models that rely on distance metrics or gradients (e.g., SVM, KNN).
2. Name techniques for handling missing data in a dataset.
   • Answer: Techniques include deletion (listwise or pairwise), imputation (mean, median, mode), prediction models (regression), and using algorithms that handle missing values (e.g., XGBoost).
3. Explain feature extraction and its importance in machine learning.
   • Answer: Feature extraction involves transforming raw data into a set of relevant features that facilitate model training and improve predictive performance by capturing meaningful information from the data.

Deep Learning and Neural Networks

1. What is deep learning, and how does it differ from traditional machine learning?
   • Answer: Deep learning is a subset of machine learning using neural networks with multiple layers to learn representations of data. It differs by automatically learning hierarchical features from data without explicit feature extraction.
2. Describe the structure of a typical neural network.
   • Answer: A neural network consists of an input layer, hidden layers (with neurons and activation functions), and an output layer. Connections between layers have weights adjusted during training to optimize performance.