Computing environment for ML4Trading

From Quantwiki
Jump to: navigation, search

Our course uses readily available Python libraries that work on Windows, MacOS and Linux. However, sometimes these libraries can be tricky to set up. In addition, the data sets we're going to work with can be somewhat large.

If you want to, you can probably set up the libraries and data on your own computer, and you are invited to do so if you like. However, we don't have the personnel to help people out on all the different platforms out there. Accordingly, we have set up a computer on campus that has all of the software and data you will need for the course.

If you want to ignore our advice and run on your own computer

We suggest, if you are in the course at GT that you follow the instructions below to use our server If after you try that, you want to install on your own computer, see the QSToolKit_Installation_Guide for more information on this.

If you are using MacOS or Unix (including Linux, RedHat, Ubuntu, Fedora)

If you want to use a Mac or Unix box without installing the software locally open a terminal window and type:

xhost +
ssh -X

And log in with your GT password. Continue with the "Configure your environment" step below.

If you are using MS Windows

Our server gekko is a Unix machine that you must use via a terminal window. In order to do this you will need Xwindows and ssh. Here are instructions:

  1. Install an Xwindows client. Georgia Tech offers a free one for students: GT OIT Software Download. Scroll down and get "X-Win32 v10 (with SSH)." We've also heard Xming is good.
  2. If you use the GT download you will also get ssh. If you did not use the GT download, you must go elsewhere to get ssh. You might try putty.
  3. Once you have Xwindows and ssh, open a terminal window and type "ssh -X" to connect.

Configure your environment

Skip to Step 3 of the installation guide QSToolKit_Installation_Guide#Step_3:_Installing_QSTK

Some hints on using Unix and Python

  • If you don't already have a file editor preference, use "nano". It is simple and easy to use.
  • Within Python use "execfile('')" to test the program you're writing.
  • Here's a cheat sheet for unix commands: [1]
  • Use two terminal windows (at least). Edit your program in one window and use the other window to run Python.
  • Use "evince" to view your graphical output, e.g.: "evince normalized.pdf"
  • Remember to reload the file on evince after you've recreated it by rerunning your program.