Applications for e-Government : global terrorism analysis
Master Thesis
Author
Kapsis, Dimitrios
Καψής, Δημήτριος
Date
2024-02Advisor
Prentza, AndrianaΠρέντζα, Ανδριάνα
View/ Open
Keywords
Big data ; Data science ; Data analysis ; Machine learning ; Neural network ; Deep neural networks ; Terrorism ; Terrorist attacks ; Terrorist groups ; e-GovernmentAbstract
Terrorism has a great impact to the society both in the people’s lives and in the economy. Many devastating terrorist attacks have occurred worldwide over the years causing extensive property damage but also leading to the loss of many human lives. Artificial Intelligence (AI) has been a very practical asset to counter this menace. In this master thesis, the Global Terrorism database (GTD) is studied, along with a second dataset QFactors_Dataset based on GTD, which includes additional features of different nature, such as the number of refugees and immigrants, average lifespan, education level, etc. Various Deep Neural Network (DNN) models are created and compared to categorize terrorist attacks based on the political identity of the terrorist groups. First, data cleaning was performed, and data preprocessing methodologies were employed to make the raw data suitable for training. More specifically, it was essential to convert our text data (regions, countries, terrorist groups etc.) to numbers, and normalize the data, in order to be able to use them in our models. The dataset consists of records for both successful and failed attacks with the use of various weapons, attack methods and targets. The models consider various factors, including, but not limited to the date of the attack, the country, the terrorist group, the number of deaths, the type of the attack along with the weapons used. The second dataset had some additional features of different nature, like the number of refugees and immigrants, life expectancy, the education level, and others. All these factors are analysed in later sections.