Exploratory data analysis
Palmer Penguins — load, inspect, missing values, plots.
Fifteen modules across two levels. ML1 covers classical machine learning; ML2 goes deeper into neural networks and PyTorch. Each section includes a PDF preview and a link to view the notebook on GitHub.
Palmer Penguins — load, inspect, missing values, plots.
Auto MPG — train/test split, metrics, coefficients.
Iris — scaling, confusion matrix, decision boundary.
Wine dataset — train a tree, plot, feature importance.
Sonar — rock vs mine classification.
Iris — DMatrix API and sklearn classifier.
XGBoost, LightGBM, CatBoost on bank marketing data.
Mall customers — elbow method, segments, silhouette.
Penguins & wine — variance, 2D projection, reconstruction.
Microsoft stock — seasonality, future horizon, MAE.
Tensors, autograd, and training a model with PyTorch.
Build and train a feedforward network in PyTorch.
Dropout, batch norm, and learning rate schedulers.
Convolutional neural network for image classification.
LSTM for sequence-based time-series prediction.