Data mining, machine learning, depth learning, referral algorithms and the relationship between the difference summary _ depth Learning

A bunch of online searches, and finally the links and differences between these concepts are summarized as follows:

1. Data mining: Mining is a very broad concept. It literally means digging up useful information from tons of data. This work bi (business intelligence) can be done, data analysis can be done, even market operations can be done. Using Excel to analyze the data and discover some useful information, the process of guiding your business through this information is also the process of data mining.

2. Machine Learning: machine learning, a cross subject of computer science and statistics, the basic goal is to learn a x->y function (mapping) to do classification or regression work. The reason why it is often combined with data mining is that many of the data mining work is done through machine learning algorithms, such as CTR estimates of ads, PB-level Click logs can get a predictive model through a typical machine learning process, which enhances Internet advertising clicks and returns. Personalized recommendation, or through a number of machine learning algorithms to analyze the platform for various purchases, browsing and collection of logs, to get a recommendation model to predict your favorite products.

3. Depth study: Deep learning, machine learning inside now compare fire of a topic (large pit), itself is the derivation of neural network algorithm, in the image, voice and other rich media classification and recognition has achieved very good results, So the major research institutions and companies have invested a lot of manpower to do related research and development.

In summary, data mining is a very broad concept, the most common methods of data mining from the machine learning this discipline, in-depth learning is the machine learning a class of fire-comparison algorithm, in essence, the original neural network.
4. Recommended algorithm: Machine learning is the method, artificial intelligence/data mining is the application, can use machine learning, can also use other methods. Data mining has a lot of application scenarios, the recommendation system is one of the business objectives of a clear, has a certain history, into the system, has formed a more perfect experience accumulation of the application scene. There are many applications in data mining that need to be developed, even if it is possible to dig out valuable patterns. Like Recommender systems, computer vision, and NLP, these values are known to be more fortunate than others. Write the Book of course everything to write, is there something in machine learning, recommended system books can not be written. Moreover, these books focus on different, recommended system of machine learning algorithm closer to the recommended business, focusing on algorithmic applications, application effects, the impact on specific business indicators, the entire system to bring limitations or upgrade and so on. Certainly not like the statistical learning basis to tell you a pile of down the process, talk about statistical characteristics, but not like statistical learning theory to give you a few proofs. These books are all about machine learning, but the angles are different. If the book of machine Learning Algorithm and the book of recommendation system is about the same as the machine learning algorithm, then the book about machine learning can also be thrown away.