Project of team 34 for the course EE558 - A Network Tour of Data Science (EPFL)

Predicting the nature of terrorist attacks

This github is the main platform for our project in the cours EE558. We are working with a dataset provided by UCSC (https://linqs-data.soe.ucsc.edu/public/lbc/TerrorAttack.tgz). The goal of this project is to to predict the labels (location or type of attack) of an attack only by using the features given in the dataset.

The main file NTDS_team34.ipnb is separeated in 4 different sections:

• section 0: creation and loading of the different objects needed for the project. The only loading is the adjacency matrix, which is provided by UCSC. This section is also where all the needed functions are created.
• section 1: this section is the pre-analysis of our data. A correlation of the features is performed as well as the distribution of the type of attack
• section 2: this section is focused on predicting the locations of attacks with the features.
• section 3: this section is focused on using the features to predict the type of attack

You should be able to have all the results by just running the main file and patiently wait in your seat! :) -> estimated run time: 10 to 15 minutes.