Machine learning, from scratch
A series that builds up the fundamentals one model at a time, starting with the math and ending with support vector machines.
Machine Learning: A High-Level OverviewA brief orientation for complete novices in the field of machine learning: what machine learning is, how it is used, and the general framework for how it works.The Math Behind Machine LearningIf terms like gradients, dot products, or partial derivatives feel a little rusty, this post walks through the key math behind machine learning: linear algebra, calculus, and probability, without diving too deep.How Linear Regression Works (and Why it's the Foundation of Machine Learning)What linear regression is, how it makes predictions, and how it learns from data. This one model introduces feature vectors, model parameters, loss functions, and gradient descent.From Lines to Layers: The Linear Perceptron and the Rise of Neural NetworksAt the heart of today's neural networks lies a surprisingly simple idea: the perceptron. How it works, how it learns to separate data, and how stacking these units leads to powerful models.Logistic Regression: Turning Probabilities into PredictionsHow logistic regression turns inputs into probabilities and predictions, how it differs from the linear perceptron, and how it powers real-world decisions like sports predictions.Margins, Kernels, and Support Vectors: An Intuitive Guide to SVMsWhat makes a good decision boundary? SVMs maximize the distance from it, and with the help of kernels they can reshape the coordinate space to handle complex data.