Installation of machine Learning-octave environment

Recently began to look at Coursera above the machine learning course, the above mentioned a software--octave, so I transferred the following blog.

Do not know what is the specific reason, I download octave-4.2.1-w64-installer.exe, the speed is extremely slow, so downloaded Octave-4.2.1-w64.zip, after decompression in accordance with http://wiki.octave.org/ Octave_for_microsoft_windows inside, open the. bat file in the folder/.vbs file, and then you can use it directly.

Download page: https://ftp.gnu.org/gnu/octave/windows/

The following is the original blog text

Octave is a programming language for interpreting classes, and it is open source software under the GNU Project, which is relatively familiar to the matlab,octave and MATLAB syntax are basically consistent, are used to quickly do some powerful matrix operations to use, The biggest difference is that octave is completely free and open source, which means that anyone can directly use or even modify the source code, but MATLAB is commercial software, and the price is very expensive, universities and research institutions will use more, and pay will have the perfect technical support, In the country is usually very convenient to get cracked version, but Octave is both free and lightweight, and with the 4.0 after the interface and graphics features have become strong, this should be more than the cumbersome to install the MATLAB to make a lot of people, and in the case of a large number of cases octave already have all the features you need

In addition octave is precisely because of the advantages of matrix operations, can be many complex need to iterate the computation into a matrix or vector to calculate, this conversion can often be reduced to dozens of lines of code to a few lines of code can be completed, and built-in numerical analysis of the performance of the library is higher than the general self-written, generally in the data analysis, Machine learning can use octave to quickly do a good model, and then put into production can be used Python,java or C + + to achieve, and the various languages have a corresponding matrix database, this also do not have to write code, so use octave can focus on the model and data analysis itself, Without considering too much about the calculation, so the efficiency is very good, said so many of the following is the specific download and installation process:

Octave's official address is: https://www.gnu.org/software/octave/after entering the first page of the download button can quickly see the source of the download, for Linux-specific distribution can be directly using the System package management to install, one step in place; You can also download the source package to compile the installation, and install general software Common, for Windows can download the installation package directly to install,

The Windows installation package is: https://ftp.gnu.org/gnu/octave/windows/

Linux Source package is: https://ftp.gnu.org/gnu/octave/

Note: For the current version of the latest is 4.2.1, it is recommended to install the 4.0.1 up version, but do not install version 4.0, 0, because this version has a serious bug, and this bug was fixed in 4.0.1

For Windows under the installation of nothing to say, download the installation package run and then click on the next step can be installed, the desktop will appear after the installation of two shortcut links, one is the CLI after the start of the command line form, one is the GUI is the user interface form, Not only can execute commands but also be used as code editors, Octave's source code is. m format, command line is fast and stable and can also draw directly, so we usually use the command line

For Linux installation is to use general compilation steps to install, note that if it is a pure command line of Linux then the drawing command and GUI are not used, do not rely on the graphical interface is not bootable, but still can use the command line most of the features, Download here is the latest 4.2.1 source package, below to start the installation:

Release the installation package and enter the source directory: tar -xvzf octave-4.2. 1. tar. GZ && cd octave-4.2. 1/

Execute ./configure--help can view all the compilation parameter descriptions, which are installed directly to the specified directory by default: ./configure--prefix=/usr/octave

If you are missing some of the dependent package error at this time, you can install some of the following packages:

Yum Install lapack64 lapack64-develyuminstall readline-devel

After installation, perform the following: ./configure--prefix=/usr/octave is normally normal.

Then compile: Make -j4

The compilation process is very long, generally in half an hour to one hours or so, patiently wait, compiled after the completion of a bunch of warnings, the general meaning is that the lack of graphics component dependencies, will not be able to use the drawing function, that is, the interface can not be used, because our system is the command line, so directly ignore it,

Perform installation: make install

After the installation of this decompression package with the same source code piece about 3G, this can be deleted, and then into the/usr/octave installation directory, the installation directory about 500M, the installation directory below the bin is the path of the octave executable file, Can be used to run the source file directly or go to the command line execution instructions

At this point the execution: ./bin/octave-cli can enter the octave command line, as follows:

　　

In addition if careful will find, execute ./bin/octave can also enter the command line and as above, in fact, this is a GUI launcher, if you have a graphical interface on the system to start the GUI interface, but is currently a pure command line system, So the execution of the first print the error, and then into the command line interface, so use the same as the above OCTAVE-CLI

The above is the simple introduction and installation of octave

Transfer from http://www.cnblogs.com/freeweb/p/7124589.html

Installation of machine Learning-octave environment