Skip to content

hw1.md

Latest commit

 

History

History
183 lines (141 loc) · 7.12 KB

hw1.md

File metadata and controls

183 lines (141 loc) · 7.12 KB

home | copyright ©2016, tim@menzies.us

overview | syllabus | src | submit | chat

Homework1

Before you begin

  • Get your URL shortener going. Go get an account at http://tiny.cc.
  • Get your Github environment going (public github, not ncstate). Add timm and bigfatnoob as collaborators to that repo.
    • The name of that repo should be fss16xxx where xxx is anything you like.
    • Add directories to that repo as follows
project
paper
read   -- all your weeklies readings go here
code/1 -- code for week1.
code/1/README.md : answers to any questions asked this week
code/2
code/3
etc
  • Get your development environment going. This should be: - code that is checked into git and saved, regularly to your Gibhub account, and shared with any team members. - Some Unix-type environment: LINUX, Mac (with xcode installed), c9.io (a free web-based IDE). - Some folks report success with Windows and gitbash (https://git-for-windows.github.io/) but that is up to you. - GitBash https://git-for-windows.github.io/ - Chocolatey package manager https://chocolatey.org/ - Note that all your team members need to running the same environment (which makes c9.io an attractive option).
  • For your project, pick a development language. Use anything you want but be aware that the lecturer and tutor can only support students using Python on LINUX and Mac. Note also that all your team members should agree on the language.
  • Review your team. Ideally, a team should contain at least one moderate expert in what ever language you are using. Got to http://tiny.cc/fss2016give and check out our random assignments to the teams. Ask for a change if NO ONE on your team is expert on the language you are using. If you make a change: - mark your people with the same letter in column C, BELOW Row 20 - add your Gibub url in column B ABOVE row 20

Read Something

Write a summary of one research paper from 2012 relating to data mining and software engineering. For more details see Reading12345678.

Install Something

Into a fresh subdirectory, download and unpack https://github.com/dotninjas/dotninjas.github.io/archive/master.zip. Important note: the pathname to that directory should not have any spaces in it.

  wget https://github.com/dotninjas/dotninjas.github.io/archive/master.zip
  unzip master.zip 
  mv dotninjas.github.io-master/* .
  cd ninja/
  sh ninja

That should generate the following lame Ascii Art version of a ninja throwing a star.

          ___                                                             
         /___\_/                                                          
         |\_/|<\                         Command line ninjas!
         (`o`) `   __(\_            |\_  Attack!                               
         \ ~ /_.-`` _|__)  ( ( ( ( /()/                                   
        _/`-`  _.-``               `\|   
     .-`      (    .-.                                                    
    (   .-     \  /   `-._                                                
     \  (\_    /\/        `-.__-()                                        
      `-|__)__/ /  /``-.   /_____8                                        
            \__/  /     `-`                                               
           />|   /                                                        
          /| J   L                                                        
          `` |   |                                                            
             L___J                                                        
              ( |
             .oO()                                                        
_______________________________________________________
/usr/bin/java
/usr/bin/gawk
/usr/bin/python
/usr/bin/zip
/usr/bin/unzip
/usr/bin/git
/usr/bin/perl
ATTENTION: missing nothing; can you install nothing?

ninja.rc v1.0 (c) 2016 Tim Menzies, MIT (v2) license

NINJA: workspace/ninja 37>

Note that list of installed files (all the '/usr/bin' stuff). If any of those are absent, do

sudo apt-get install XXX

Run Something

For the following commands, first

 type command
 command

e.g.

 type eg0
 eg0

Hand in one file with 5 lines of text describing each eg.

The commmands are

eg0 eg1 eg2
eg3 eg4 eg5
eg6 eg7 eg8
eg9 eg10

Notes:

  • Hint. Read the notes in https://github.com/dotninjas/dotninjas.github.io/blob/master/ninja/ninja.rc.md
  • To understand eg5, look up this paper and read the definitions column2, page1.
  • For eg6, answer the following question. Is this code doing stratified cross-validation (read more here.
  • For eg8, what is a named column?
  • For eg9, note that eg8 has to be run first.
    • What is the advantage (if any) of separating (a)the reporting of a data mining run (as done in eg9) from (b)the execution of that run (as done in eg8)?
    • When do shells scripts like
  • For eg10, write 2 lines describing any two of the following learners j48, jrip,nb, rbfnet bnet
    • Browse ninja.rc till you find the definition of that function; e.g. rbfnet
    • Look up its class name; e.g. RBFNetwork
    • Go to http://wiki.pentaho.com/display/DATAMINING/RBFNetwork and search (top right).
    • Go to the package definition (usually, first search return).
    • Read!
    • By the way, for some cool notes on bnet, see here

Working with Cloud9

As of June 2015, the procedure for doing that was:

  • Go to Github and create an empty repository.
  • Log in to Cloud9 using your GitHub username (at http://c9.io, there is a button for that, top right).
  • Hit the green CREATE NEW WORKSPACE button
    • Select Clone from URL;
    • Find Source URL and enter in http://github.com/you/yourRepo
    • Wait ten seconds for the screen to change.
    • Hit the green START EDITING button.

This will drop you into the wonderful Cloud9 integrated development environment. Here, you can edit code and backed up your code outside Cloud9, over at Github.com (which means that if ever Cloud9 goes away, you will still have your code).

The good news about Cloud9 is that it is very easy to setup and configure. The bad news is that each Cloud9 workspace has the same limits as Github- a 1GB size limit. Also, for CPU-intensive applications, shared on-line resources like Cloud9 can be a little slow. That said, for the newbie, Cloud9 is a very useful tool to jump start the learning process.

For sites other than Cloud9, see Koding, Nitrous.IO and many more besides.