Stanford Machine Learning notes（Lec 1）
Problem definition：
Arther Samuel(1959)
Machine learning problem:field of study that gives computers the ability to learn without being explicitly programmed.
Tom Mitchell(1998)
Well-posed Machine Learning: A computer program is said to learn from experience E with respect to some task T and some performance measure P, if it's performance on T, as measured by P, improves with experience E.
Example: Samuel let the computer to learn how to play chess by playing chess against itself and it finally played much better than Samuel himself.
Four main parts of this course:
1.Supervised learning: Regression problem for continuous variable prediction
Classification for discrete variable prediction
Several right answers are given, and algorithm knows how to judge by learning from them.
2.learning theory
3.Unsupervised learning: for example, clustering, no right answers are given.
4.Reinforcement learning
That is to make a sequence of decisions. For example, teach a helicopter how to fly.
