1 Classes of Problems in Machine Learning

Back to Data-Science

Formal Definition: program learns from experience E with respect to class of tasks T and performance measure P, if performance (P) at tasks in T improve with experience (E)

Supervised Learning

Unsupervised Learning

Linear Regression One Var

In regression problems, given input variables, try to find a continuous expected result function

Gradient Descent

Recall that the gradient vector ∇f = (fx, fy) is like a derrivative for a field (fx = ∂f/∂x)