This project is about to classify if an event is terrorist or other forms of crime as per the Global Standards.
-
Exploratory data analysis
-
Statistical tests to check the significance of variables
-
Treating the missing data
-
Backward Elimination
-
Modelling with null value imputation and by dropping nulls
-
Crossvalidating the models
-
Principal Component Analysis,K means Clustering
-
Synthetic Minority Over-sampling Technique
-
Evaluation of the models
-
Application of the model and advantages of it
Compared to most types of criminal violence, terrorism poses special challenges to a nation and exhausts all of its resources in prevention of it including the loss of life. While the human cost is devastating, the economic impact may be larger than most realize. Terrorism is one of the parameters that tourists check for, before visiting a country and hence if a Nation has more prevalent terrorism, chances are, despite its fascinating tourist attractions, it might end up in little to no Tourism. In response, there has been growing interest in researching about terrorism, their motives and most vulnerable target groups that are attacked. One thing that is infrequent is one common definition of Terrorism throughout the world. This is why the Global Terrorism Database has also included those events here which don’t confirm to global inclusion criteria for terrorism but are identified as terrorist events by the locals. Our aim in this project was to classify the events as terrorist or other forms of crime based on the Global inclusion parameters so that we can help the various intelligence agencies to drill down their study exclusive to Terrorism, avoiding any ambiguity that would come with the raw data(G.T.D).
In the course of this project i use Supervised learning classification. Classification is a subcategory of supervised learning where the goal is to predict the categorical class labels (discrete, unoredered values, group membership) of new instances based on past observations.
More information about the project is in the attached Final Report.
Since the data is big and a large amount of insights can be drawn from it,i have used tableau for the Exploratory data analysis and generating insights. Below is a peak into the tableau visualisation of different perpetrator groups of different attacks:
Complete visualisations can be viewed here: https://rb.gy/qzj7co
I have also made separate visualisations for the top 5 most adversely affected countries and made a forecast of attacks that can happen there till 2025.This can be viewed here: https://rb.gy/jh3sri