"Machine Learning Basics" machine learning Cornerstone Course Learning Introduction

What is machine learning?

"Machine learning" is one of the core research fields of artificial intelligence, its initial research motive is to let the computer system have human learning ability to realize artificial intelligence.
In fact, since "experience" is mainly in the form of data in the computer system, machine learning needs to try to analyze the data, which makes it become one of the innovative sources of intelligent data analysis technology.
Machine learning is a way of building complex systems, and perhaps relying on our brains to write all the rules that deal with a problem may not be easy to do, then we let the machine do its own learning based on data and materials and deal with the problem on its own.

The relationship between machine learning and data mining

Data mining can be considered as a non-trivial process to identify effective, novel, potentially useful, and ultimately understandable patterns in huge amounts of data . In fact, as the name implies, data mining is trying to find useful knowledge from the massive data. In general, data mining can be seen as a cross between machine learning and databases, which mainly uses the technology provided by the machine learning community to analyze massive amounts of data and use the technology provided by the database community to manage massive amounts of data.
In a word, machine learning is a direction of artificial intelligence, and machine learning is a way to realize data mining.

The relationship between machine learning and statistics

Statistics is the activity of using data to speculate on an unknown process, which is a way to realize machine learning.
Traditional statistics pay more attention to mathematical inferences, rather than machine learning techniques that focus on how the computer calculates the results of inferences.

Application Scenarios for machine learning

• When we are not able to implement the system by manual programming, that is impossible to exhaust all the rules of the case
• When we can't easily define the answer
• When a quick decision is needed, it is not possible for humans to do so, such as the high-frequency trading of the stock market
• When targeting a wide range of user personalization services, such as marketing strategies for consumers

The key to machine learning

1. (pattern) There is a potential pattern or rule that can be learned
2. (Definition) is not easy to implement programmatically
3. (data) has information about a pattern

The actual definition of machine learning
Machine learning is actually defined

As shown, since we do not know F, we then put the data on the hand D, through the machine learning algorithm A, to get a recommended function g, we want G and the ideal target function f closer to the better.
We can put a lot of possible formulas into a set of "hypotheses" H (which includes all possible g), what the machine learning algorithm needs to do is to choose from h the hypothesis that it feels best, that is, G.
Here, we give a clearer definition of machine learning: machine learning is, we start with the data, through the machine learning algorithm, to work out a hypothesis g, I hope G can be very close to our desire for the ideal model F.

Next

Next, I will follow the course of online learning to organize the daily learning of machine learning algorithms, and practical problems, do more experiments, and strive to get a better learning effect, I will be firm belief, more efforts to catch up with the pace of excellence.

Reprint please indicate the author Jason Ding and its provenance
GitHub home page (http://jasonding1354.github.io/)
CSDN Blog (http://blog.csdn.net/jasonding1354)
Jane Book homepage (http://www.jianshu.com/users/2bd9b48f6ea8/latest_articles)

"Machine Learning Basics" machine learning Cornerstone Course Learning Introduction