Course04 · Machine Learning

1. Day 141 What Machine Learning Is and Is Not complete
2. Day 142 Supervised, Unsupervised, and Reinforcement Learning complete
3. Day 143 The Machine Learning Workflow complete
4. Day 144 Train, Validation, and Test Splits complete
5. Day 145 Overfitting and Underfitting complete
6. Day 146 Your First Model with scikit-learn complete
7. Day 147 An End-to-End Classification Exercise complete

Project: First End-to-End Model — Train, evaluate, and document a classifier on the iris dataset with a proper train/test protocol and an honest error analysis.

1. Day 148 Linear Regression complete
2. Day 149 Loss Functions and Least Squares complete
3. Day 150 Multiple and Polynomial Regression complete
4. Day 151 Regularization: Ridge and Lasso complete
5. Day 152 Regression Metrics complete
6. Day 153 Linear Regression from Scratch complete
7. Day 154 A Complete Regression Project complete

Project: House Price Model — Build a regularized regression model on a housing dataset with feature analysis and residual diagnostics.

1. Day 155 Logistic Regression complete
2. Day 156 Decision Boundaries complete
3. Day 157 k-Nearest Neighbors complete
4. Day 158 Naive Bayes and Text Classification complete
5. Day 159 Precision, Recall, ROC, and Choosing Thresholds complete
6. Day 160 Class Imbalance complete
7. Day 161 A Complete Classification Project complete

Project: Spam Classifier — Build a text spam classifier with proper handling of class imbalance and a precision/recall trade-off analysis.

1. Day 162 Decision Trees complete
2. Day 163 Random Forests complete
3. Day 164 Gradient Boosting complete
4. Day 165 XGBoost and LightGBM in Practice complete
5. Day 166 Hyperparameter Tuning complete
6. Day 167 Cross-Validation Done Right complete
7. Day 168 Winning on Tabular Data complete

Project: Tabular Challenge — Compete against your own baseline on a tabular dataset: tuned gradient boosting versus a linear model, with a documented comparison.

1. Day 169 Support Vector Machines complete
2. Day 170 Feature Scaling and Encoding complete
3. Day 171 Feature Engineering complete
4. Day 172 Feature Selection complete
5. Day 173 scikit-learn Pipelines complete
6. Day 174 Handling Missing Data complete
7. Day 175 Features Beat Algorithms complete

Project: Feature Engineering Challenge — Improve a fixed model’s performance purely through feature work on a raw dataset, documenting each feature’s measured impact.

1. Day 176 Choosing the Right Metric complete
2. Day 177 Learning Curves and Diagnostics complete
3. Day 178 Interpreting Models: Importances and SHAP complete
4. Day 179 Fairness and Bias in Models complete
5. Day 180 Data Leakage complete
6. Day 181 Baselines and Error Analysis complete
7. Day 182 Writing a Model Report complete

Project: Model Audit Report — Audit a trained model for leakage, fairness, and failure modes, and write a decision-ready model report.

1. Day 183 Clustering with k-means complete
2. Day 184 Hierarchical Clustering and DBSCAN complete
3. Day 185 Principal Component Analysis complete
4. Day 186 t-SNE and UMAP complete
5. Day 187 Anomaly Detection complete
6. Day 188 Recommender Systems complete
7. Day 189 A Segmentation Study complete

Project: Customer Segmentation — Cluster a customer dataset, reduce it for visualization, name the segments, and defend the number of clusters chosen.

1. Day 190 The ML Project Lifecycle planned
2. Day 191 Building Datasets and Labeling planned
3. Day 192 Time Series Forecasting Basics planned
4. Day 193 Saving and Versioning Models planned
5. Day 194 Serving a Model over an API planned
6. Day 195 Monitoring Models in Production planned
7. Day 196 Section Project: An ML Service planned

Project: Section Project: Deployed ML Service — Train, persist, and serve a model behind a FastAPI endpoint with input validation, tests, and a monitoring plan.