Global Terrorism Predictive—Analysis

Abstract

In the recent view of increasing number and lethality of terrorist attacks, it has become important for us to recognize a strategic vision to help prepare and prevent such events from happening. This paper includes descriptive and predictive analyses of Global Terrorism Database which reveal vital information about the trends of such events and help identify the perpetrators of any such future terrorist activities. The descriptive phase covers elucidation of the dataset to identify useful features for forecasting and predictive phase involves data manipulation and compares the performance of various multi-class classification and regression algorithms like decision trees, random forest, etc., on the dataset. Python with Scikit-learn library was used for the experimentation purpose.