Skip to content

Code from the United States Agency for International Development (USAID) and Humanity United (HU) challenge for Atrocity Prevention Modeling in collaboration with NASA’s Center of Excellence for Collaboration (CoECI).

License

Notifications You must be signed in to change notification settings

NASA-Tournament-Lab/CoECI-USAID-Atrocity-Prevention-Model

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

CoECI-USAID-Atrocity-Prevention-Model
=====================================

Code from the United States Agency for International Development (USAID) and Humanity United (HU) challenge for Atrocity Prevention Modeling in collaboration with NASA�s Center of Excellence for Collaboration (CoECI).

This package contains the top 5 solutions submitted for the "Tech Challenge for Atrocity Prevention" competition.

You can find the competition task problem statement here: http://community.topcoder.com/longcontest/?module=ViewProblemStatement&rd=15761&pm=12634.

It is necessary to read the problem statement in order to understand what the solutions are doing.

In order to use the package, compile main.cpp file as follows:

g++ --std=c++0x -W -Wall -Wno-sign-compare -O2 -s -pipe -mmmx -msse -msse2 -msse3 -Wl,--stack,4194304 -o main main.cpp

The obtained executable is to be run with the following 6 parameters:

main <input folder> <output folder> <solution> <first training day> <first testing day> <last testing day>

<input folder> must contain all data files. The archive with all data can be downloaded at <PUT URL HERE>.

<output folder> is the folder where the results must be saved. For each day D of the testing period, the file named res_D.txt will be created in the folder. It will contain 3671 numbers -- the confidences that an atrocity
will occur within next 30 days in each of the 3671 regions.

<solution> is the number of solution to be used. It can be 1, 2, 3 or 5 for 1st, 2nd, 3rd or 5th placed solution, respectively. The places were awarded based on a well-defined objective metric, but still each solution has its stronger and weaker sides. In particular, it's known that 1st and 2nd place solutions complement each other pretty well.

<first training day>, <first testing day> and <last testing day> define the training and testing periods. You can read the problem statement for more details.

Three settings that correspond to example, provisional and system test of the actual competition are as follows.

Example test: <first training day> = 11284, <first testing day> = 15302, <last testing day> = 15636.
Provisional test: <first training day> = 11284, <first testing day> = 15667, <last testing day> = 16367.
System test: <first training day> = 11284, <first testing day> = 16398, <last testing day> = 17644.

For example, if you want to execute 2nd place solution on provisional test, take input data from "data" folder and save results to "out" folder, here is the command to do so:

main data out 2 11284 16398 17644

Finally, please note that 4th placed solution is implemented in Java (while all other solutions are implemented in C++), so it needs to be executed in a different, but a very similar way: java -jar place4.jar <input folder> <output folder> <first training day> <first testing day> <last testing day>

The data is currently located in a Google Drive folder but will be uploaded to the appropriate Github repository for this project. At this time, please access the data using this link: https://docs.google.com/file/d/0B5LAtOowzIeoNjNMZ1VZTzZDNTg/edit?pli=1.

About

Code from the United States Agency for International Development (USAID) and Humanity United (HU) challenge for Atrocity Prevention Modeling in collaboration with NASA’s Center of Excellence for Collaboration (CoECI).

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 3

  •  
  •  
  •