ML vs. Traditional Programming, Supervised, Unsupervised, and Reinforcement Learning, Real-world analogies for each paradigm

Healthcare: Disease prediction, Finance: Fraud detection, E-commerce: Recommendation systems, Autonomous vehicles

Structured (tabular) vs. Unstructured (images, text), Labeled vs. Unlabeled data, Common dataset sources (Kaggle, UCI)

End-to-end workflow, Data Collection → Preprocessing → Model Training → Evaluation → Deployment, Iterative nature of ML projects

Handling missing values (Mean/Median imputation), Dealing with outliers (IQR method, Z-score), Duplicate detection and removal

One-Hot Encoding for categorical data, Feature Scaling: Standardization (Z-score) & Normalization (Min-Max), Creating polynomial features

Train-test split (80-20 rule), Stratified sampling for imbalanced data, K-Fold Cross-Validation (k=5, k=10)

Line of best fit and cost function (MSE), Gradient descent optimization, Implementation: scikit-learn's LinearRegression

Logistic Regression (sigmoid function), Decision Trees (Gini impurity, entropy), Random Forest (ensemble methods)

Classification metrics: Accuracy, Precision, Recall, F1-Score, Confusion matrix interpretation, Regression metrics: RMSE, R² Score

K-means clustering (elbow method), Hierarchical clustering (dendrograms), DBSCAN for density-based clustering

Principal Component Analysis (PCA), t-SNE for visualization, Feature importance analysis

Customer segmentation (RFM analysis), Anomaly detection (isolation forest), Market basket analysis