
Already know some of this? A 2-minute test skips the lessons you know.
Take the placement testA free 23-lesson course built from freeCodeCamp.org's 3h 54m of YouTube video. 3 days at 90 minutes a day, with study notes and a quiz in every lesson. A 2-minute placement test skips what you already know.
Taught by freeCodeCamp.org on YouTube. Videos play in YouTube's own player, so the creator keeps every view.
Introduces the course and instructor, then loads the Magic Gamma Telescope dataset into a Google Colab notebook with pandas.
Defines AI, ML, and Data Science; explains supervised, unsupervised, and reinforcement learning; details feature types (nominal, ordinal, quantitative) and encoding methods.
Explains data splitting for generalization, the training feedback loop, validation vs. test sets, and L1, L2, and binary cross-entropy loss functions.
Covers accuracy calculation, then visualizes feature distributions by class using normalized histograms to inspect class separation.
Splits data into train/validation/test sets (60/20/20), shuffles samples, and applies StandardScaler to normalize feature scales.
Derives Bayes rule using a disease testing example, explains the naive independence assumption, MAP estimation, and implements GaussianNB.
Contrasts linear regression with logistic regression, derives the sigmoid function from log-odds, and implements LogisticRegression.
Explains maximum-margin hyperplanes, support vectors, the kernel trick for non-linear separation, and implements SVC.
Covers neurons, dense layers, activation functions (ReLU, sigmoid, tanh), gradient descent, backpropagation, learning rate, and TensorFlow/Keras model definition.
Trains a binary classification neural net with validation split, plots loss/accuracy curves, defines a train_model function with dropout, and runs a grid search over nodes, dropout, learning rate, and batch size.
Introduces regression with the Seoul bike-sharing dataset, explains residuals, RSS, TSS, R-squared, adjusted R-squared, MAE, MSE, and RMSE.
Preprocesses bike data (drops columns, filters hour), visualizes feature relationships, scales features, splits data, fits LinearRegression and a TensorFlow neural net, and compares predictions via true-vs-predicted plots.
Explains K-means clustering algorithm steps: random centroid initialization, distance assignment (expectation), centroid recomputation (maximization), and convergence.
Describes PCA as variance-maximizing projection for dimensionality reduction, contrasts with linear regression, and explains the link to minimizing projection residuals.