Wu Enda study notes
I. Introduction
1.1 Welcome
1.2 What is machine learning?
- Introduced the initial concept of machine learning-a person who is not proficient in chess wrote a program, the program played 10,000 chess with himself, and recorded each experience. Finally, the program's chess skills surpassed the programmer.
- The most recent definition-a program is considered to be able to learn from experience E, solve task T, and achieve performance metric P, if and only if, after experience E, after P evaluation, the performance of the program in processing T is somewhat Promote.
1.3 Supervised learning
Supervised learning refers to the process of adjusting the parameters of the classifier with a set of samples of known categories to achieve the required performance. It is also called supervised training or learning with a teacher.
In fact, it is a labeled training set, which achieves the best fitting effect through a certain algorithm
- Linear regression (continuous)
- Logistic regression (discrete)
1.4 Unsupervised learning
There are often such problems in real life:
(1) Lack of sufficient prior knowledge, so it is difficult to manually label categories;
(2) The cost of manual category labeling is too high.
Naturally, we hope that computers can (partially) complete these tasks for us, or at least provide some help. Common application backgrounds include:
(1) Select some representative samples from a huge sample set and label them for the training of the classifier.
(2) First, all samples are automatically divided into different categories, and then these categories are labeled by humans.
(3) In the absence of category information, look for good features.
Such an operation is actually unsupervised learning
- Usually use clustering method to achieve