SQL Python R SAS deep learning experience

SQL: Personally think that SQL is the necessary language for data work, SQL language in the common database is basically universal, learned to use one.

Reprint Please specify source: http://www.cnblogs.com/SSSR/p/7016660.html

Sas:

SAS is a good language for learning statistics, why do you say so? Because many of the current statistical books are implemented with SAS, from the principle to the realization there are a lot of books to read, as well as SAS company such a good company to support.

But SAS is really not good at learning, his programming ideas and other languages are a little different, many people compare macros to other language functions, but I think it is absolutely not so compared, because the macro is used to generate code of the program.

Regarding the study suggestion, I thought learns the SAS first, then unifies the SAS study statistics, must not study together, otherwise the difficulty will increase many.

One of the difficulties with SAS is that there is a lot of detail, and this needs to be personal, but it's easy to look at SAS help.

The PDV mechanism of SAS and the execution mechanism of macros must be understood.

SAS has a great advantage, the standard of unification, as long as the learning to be able to swim throughout the system.

R VS python:

In contrast, R is statistically much stronger than Python because Statsmodel does not give force, and new statistical methods Python cannot keep pace. In the area of data mining, Python's sklearn is stronger than R, but R's data mining is more powerful than python.

Python and R are quite the same in terms of data cleansing, but Python handles data volumes and speeds faster than R. Python is a great way to learn more about deep learning, and now most deep learning libraries support python interfaces, with very little support for R. In terms of drawing, the two are also

The generations, all have their own killer. In automation: Python is a victory over R. On the system side: Python can develop Web sites, form programs, and R is weak in this regard (shiny server requires a fee).

R vs SAS

SAS in the advantages of R is enough statistics, software standards Unified, more abundant information documents. The advantage of R is open source, low learning cost.

Deep Learning:

Now deep learning is very fire, along with the fire of an artificial intelligence, incredibly there is advertising that the AI era, the big Data era is not a few years, is it past? I personally think that at present, because the depth of neural network can not evaluate the importance of features, so it should be full-featured data is better (film, sound, video, text, etc.).

There is a lot of 0 basic learning in depth study of the remedial class, we must not believe, do not believe, if it is really so simple, why these people do not need deep learning to start a business to earn money, but also open a training course to make money? If it's that simple, how many people are there around you? If it's that simple, why do they charge so much? Careful and cautious!

SQL Python R SAS deep learning experience