Coursera Machine Learning Study notes (i)

Before the machine learning is very interested in the holiday cannot to see Coursera machine learning all the courses, collated notes in order to experience repeatedly.

I. Introduction (Week 1)

-What's machine learning

There is no unanimous answer to the definition of machine learning.

Arthur Samuel (1959) gives a definition of machine learning:

Machine learning is about giving computers the ability to learn without explicit programming.

Samuel designed a chess game in which he let the program play chess with himself and record the possible results of the position on the board. After nearly tens of thousands of, the computer learns which position on the board can be won, and ultimately overcomes the program's designer.

Tom Mitchell (1998) gives a more modern and formalized definition:

For a task T and a measure of performance p, if a computer program completes task T, the p to measure performance increases with experience E, we say that the computer program can learn by E.

For chess games, playing chess game is the task T, the game win or lose is the performance P, a game of one inning is the process of experience E.

For message classification, spam and non-spam classification is the task T, the correct rate of classification is performance p, check whether the mailing label is garbage or non-spam is experience E.

For machine learning algorithms can be divided into:

-Supervised learning

-Non-supervised learning

Some examples of machine learning applications:

-Data Mining

-Some applications that cannot be implemented by manual programming: Natural Language processing, computer vision

-some self-service programs: Referral system

-Understand how humans learn

Coursera Machine Learning Study notes (i)