Machine Learning Foundation--anaconda Environment

　　Anaconda is a Python release for scientific computing that supports Linux, MAC, and Windows systems, and contains numerous popular scientific calculations and data analysis Python packages. In addition, Anaconda provides package management and environment management capabilities to easily resolve multiple versions of Python coexistence, switching, and a variety of third-party package installation issues. Anaconda uses tools/command Conda for package and environment management, and already contains Python and related tools.
Anaconda is more of a Python environment than an IDE. Anaconda contains NumPy, pandas, Matplotlib and other libraries, so the use of Anaconda can let us avoid too much energy to spend on the environment to build, and thus quickly into the exploration of machine learning.

Anaconda's official download address is https://www.continuum.io/downloads/, but the domestic download is too slow, then we can go to Tsinghua University open source software mirror station download Anaconda installation package. In addition, the Tsinghua University TUNA team also provides a mirror image of the Anaconda Warehouse, running the following command:

Conda config--add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/
Conda config--set show_ Channel_urls Yes

You can add Anaconda Python free warehouse and run Conda install NumPy test it.

I have downloaded the Windows 64bit version of the Anaconda, the installation process is very simple, double-click the exe file, according to the prompts to click "Next" can be.

Anaconda contains more than 720 Python and R packages on science, data, engineering, and data analysis, and this version I installed requires 1.8GB of space. If you don't want to install too many packages on your computer, you can also use Miniconda, which contains the Conda, its dependencies, and the Python package.
After installation, open Anaconda Prompt and enter Conda list to list all installed packages as shown in the following figure.

As you can see, some of the libraries that we will need for our next machine learning are already installed, including: NumPy: A scientific Computing library that encapsulates a number of matrix-processing functions; Pandas: Data analysis and processing library, encapsulates a lot of data processing functions; matplotlib: Data Visualization library, Encapsulates a number of visualization functions; Scikit-learn: Machine learning Library, which encapsulates the functions of machine learning and modeling.

In the end, Conda can be understood as a tool and an executable command whose core functions are package management and environmental management. Package management is similar to the use of the PIP, and environmental management allows users to easily install different versions of Python and can switch quickly. Anaconda is a packaged collection of Conda, a version of Python, numerous packages, scientific computing tools, and so on, which is also known as a release of Python. In fact, there are Miniconda, as the name suggests, it contains only the most basic content--python and Conda, as well as the relevant must rely on, for the space requirements of users, Miniconda is a choice.

"Reference reading"
http://python.jobbole.com/86236/
Http://www.360doc.com/content/16/1029/18/25664332_602357786.shtml