Machine Learning notes, MOOC, w1_introduction

1. Definition of supervised learning and unsupervised learning

Idea of supervised learning is to teach computer how to do something.

Idea of unsupervised learning is to let the computer learn by itself.

Reinforcement Learning: Reinforcement Learning

1. Supervised Learning

We give the system a set of dataset, which is a set of "right answers ".

It is also called regress problem: predict a continuous value output.

Classification problem: predict a discrete value.

We cocould identify a problem whether is a classification one or a regression one by the output value. if it is a discrete value, it shocould be a classification problem; otherwise, it will be a regression one.

1. Unsupervised learning

We are given the data without labels. we are not told to do prediction work. Just the dataset, we are asked to find the structure of the dataset.

Clustering problem: kind of similar to the classification problem in supervised learning, the difference is that in unsupervised learning,There is no "correct answer" SetThat can be used for studying.