I. Overview
Writing the machine learning code in Python is fairly straightforward, because there are a lot of machine learning libraries under Python. The following three library numpy,scipy,matplotlib,scikit-learn are commonly used combinations, namely scientific computing packages, scientific toolset, drawing Toolkit, machine learning toolset.
NumPy: Mainly used to do some scientific operations, mainly the operation of the Matrix. NumPy brings true multidimensional array functionality to Python and provides a rich library of functions to handle these arrays. It will be used in the mathematical functions of the array, so that these mathematical functions can be directly in the operation of the array, the need to be at the python level of the loop, put in the C language operations, obviously improve the program's operation speed. "1"
SciPy: Mainly a set of scientific tools, a set of signal processing tools (such as linear algebra using a lapack library, a fast Fourier transform using a fftpack library) and some tools for numerical calculations (ordinary differential equations are solved using odepack libraries, nonlinear equations and minimum value solutions, etc.) "1".
Scikit-learn: There are a lot of machine learning related algorithms (such as clustering algorithms, SVM, etc.). "2"
Matplotlib: is a drawing tool similar to drawing engineering in MATLAB.
Two. Installation
To switch to the root administrator user, know the following command to install:
Yum-y install gcc gcc-c++ numpy python-devel scipy
Yum Install Python-matplotlib
Yum Install Python-pip
Pip Install-u Scikit-learn
CentOS Installation Machine Learning Package