Originally, I thought I didn't need to explain this question. What is the difference between data mining, machine learning, and artificial intelligence (AI), but a few days ago, because a student asked me, I thought about it. Wanting to find out that I couldn't answer it, I checked this question on Zhihu and blogs, and found that no one has written a more detailed and convincing comparison and explanation. Then I will try to talk about the difference between these based on the books and papers I have read before, as well as the communication with my supervisor. After all, a good definition can play a great role in future learning and communication. . Also fill in the relationship between data science and business analytics. The ability is limited, if there is any omission, please forgive and correct.
introduction
This article is mainly divided into two parts. The first part explains the difference between data mining (data mining), machine learning (machine learning), and artificial intelligence (AI). The difference between these three is mainly due to different purposes, and their means (algorithms, models) overlap greatly, so it is easy to confuse them. The second part focuses on the relationship between the above skills and data science, as well as the relationship between data science and business analytics. In fact, data scientists themselves are an extension of business analysts in the era of big data.
Data Mining vs. Machine Learning vs. Artificial Intelligence
Data mining: Patterns and models that purposefully extract data from existing big data
Keywords: schema extraction, big data
Data mining is to extract data patterns and models from existing information, that is, to select the most important information for future data use in machine learning and AI. Its core purpose is to find relationships between data variables. The main reason for its development is the development of big data. Traditional data analysis methods have been unable to process so many seemingly irrelevant data. Therefore, data mining technology is needed to extract the relationship between various data and variables. interrelationships to refine the data.
Data mining is essentially the basis of machine learning and artificial intelligence. Its main purpose is to extract a superset of information from various data sources, and then combine this information to make you discover that you have never Thought patterns and interrelationships. This means that data mining is not a method for proving hypotheses, but a method for constructing various hypotheses . Data mining cannot tell you the answers to these questions, it can only tell you that A and B may be related, but it cannot tell you what is the relationship between A and B.
Of course, data mining uses a lot of machine learning algorithms, but its specific context and purpose are not quite the same as machine learning.
Machine learning: Automatically learn new knowledge from past experience.
Keywords: automation, self-optimization, prediction, training data required, recommender system
Machine learning is actually a very important part of artificial intelligence, because at present, in practice, most tasks handled by artificial intelligence are actually completed by machine learning. Machine learning can learn automatically with programs and algorithms that, as long as they are designed, can optimize themselves. At the same time, machine learning requires a certain amount of training data set (training data set) for building "knowledge" from past experience.
And the most important function of machine learning in practice today is predicting outcomes. For example, machine learning has finished learning, and now there is a new data set x, and its classification needs to be predicted, and the machine learning algorithm will match the learned "knowledge" based on this new data (in fact, knowledge refers to the learned "knowledge". Mathematical model), and then classify this dataset x into a certain class C. More common machine learning, such as Amazon's recommendation system.
Artificial Intelligence (AI): A broad concept that essentially uses data and models to provide solutions to existing problems.
Keywords: deal with problems like people, a collection of technologies
Artificial intelligence is a relatively different concept from machine learning and data mining. The purpose of artificial intelligence is to create intelligent computers (don't know how to translate it well, it can be assumed to be a robot). In practice, we want this computer to be able to handle a task like an intelligent human being . Therefore, in theory, artificial intelligence includes almost everything that machines can do, including data mining and machine learning, as well as monitoring and process control.