Machine learning-supervised learning and unsupervised learning

Stanford University's Machine learning course (The instructor is Andrew Ng) is the "Bible" for learning computer learning, and the following is a lecture note.

First, what is machine learning

Machine learning are field of study that gives computers the ability to learn without being explicitly programmed.

In other words, machine learning does not need to make a concrete model, but rather to allow the computer to train its own model based on a large amount of data, such as CFD software, which is based on physical models, such as transport equations.

Ii. supervised learning (supervised learning)

Each sample in the dataset has a corresponding "correct answer", which makes predictions based on these samples, divided into two categories: regression and classification.

(1) Regression problems

For example, to predict a house price, to fit a continuous curve according to the sample set

(2) Classification problems

For example: judging benign or malignant according to tumor characteristics, the result is "benign" or "malignant", is discrete

Ii. Non-supervised learning (unsupervised learning)

Unsupervised learning data sets are different from supervised learning, without any label, i.e. no corresponding "right answer". Some structure of data can be obtained from the data set by unsupervised learning, which may be divided into two different clusters, called Clustering algorithm.

For example:

Machine learning-supervised learning and unsupervised learning