Predicting-Terrorist-Attack-Using-Machine-Learning

To Reproduce Our Results

Change you working directory to the folder where you want the project. Clone the repository to your local with:

$ git clone git@github.com:ZIYU-DEEP/Predicting-Terrorist-Attack-Using-Machine-Learning.git

NOTE: the cloning process might be slower than expected as the repo contains cross-validation predicted probabilities to give users a warm start.

Then, run one of the following:

• Windows

$ chmod u+x run.sh
$ run.sh

• Linux

$ chmod +x script.sh
$ ./run.sh

NOTE: users can alter input parameters in the shell script to change the behavior of the scripts for Feature Engineering (featureEngineering.py) and Training (train.py). You can use:

$ python featureEngineering.py -h
$ python train.py -h

to check what's user inputs are available for each script. Remember that --start_clean=0 is mandatory for a warm start. However, if you changed the random seed or the hyperparameter grid, please change it to --start_clean=1 to obtain the best classifiers under the modified context.

To retrieve the best models, the recommended model and their test performances, please run the notebook under this link.

Final Project Submission Check List

• Final Project Report;
• Code Files and Notebooks;
• Train-Test Pair Description;
• List of Features;
• Feature Importances of the Recommended Model;
• Final Lists of Predicted "positive" Entities