Terroristic Tendency Detection from Human Behavior on Social Media

Abstract—Due to the increase of the number of users on the web, it becomes easier to exchange and analyze their data. This system provides a way about how to detect the tendency of terrorism on social web mainly Twitter. Jihadist groups like ISIS are spreading online propaganda using various forms of social media such as Twitter and YouTube. One of the most common methods of stopping these groups is to suspend (temporarily block) accounts that spread propaganda when they are discovered. However, this method requires human analysts to manually scan and analyze a massive amount of information on social media. This system provides a way to do this automatically (detect radical content that is released by jihadist groups) on Twitter. The main idea of this system is to predict whether a person has the tendency to be a terrorist or not. Machine learning (ML) and Natural Language processing (NLP) approaches are used to get a preliminary determination of the terrorism tendency by analyzing public tweets published on Twitter. In this system, three machine learning algorithms are used (Support Vector Machine (SVM), Naive Bays and AdaBoost), the Accuracy of the three algorithms are compared to decide which one is better when using our data-sets. As a conclusion our system aims to reduce the increase of of distributing terrorism on the social media (Twitter).

Finally will discuss how our system can be modified to work on other behaviors, and to contribute in other manners in the human behavior researches. Keywords—Natural Language Processing, Machine Learning, Behavior analysis, Social media analysis, Twitter, Facebook , YouTube.

I. INTRODUCTION SINCE the late 1980s, the Internet has become a central of communication between all people overall the world, more and more people around the world are enjoying the benefits that the Internet provides. There many technologies have been developed for reconnecting people across continents, in 2014 there were more than 4 billion Internet users and this number is still growing [26]. Internet technology comes with numerous benefits including the increasing ease of sharing, fast accessing information and idea. Businesses, consumers, organizations and governments started using internet as a medium for communicating with each other. It also created a perfect atmosphere for various terrorist organizations to spread radical information (in a major way, bigger than ever before) that aid their own twisted causes. Federal officials claim that the Islamic State (ISIS) and other similar extremist organizations are currently engaged daily in the attempt of utilizing social media and the Internet to recruit and radicalize dissected citizens.

Thus, it is crucial to detect such groups by scanning and then analyzing the streams of data coming from social networks. There are many terrorist organizations in the world and almost everyday news outlets reports that a terrorist attacks in different regions of the world. According to [20], terrorism has the following characteristics:

• It involves violent or dangerous acts to human life that violating the law (be it federal or state).
• Seem intended to intimidate a civilian population
• Intends to violate the governmental policies by intimidation or coercion
• They use mass destruction, assassination or kidnapping as a means to an end to affect the conduct of a government.

Most terrorist groups use the Internet to spread propaganda and their radical information [1][2]. Which usually includes audio , video, virtual messages and presentations that contain caption and justifications of terrorist activities (sometimes even promotion of terrorist activities). The aim of this propaganda is influence of opinions, emotions and attitudes. Another propaganda major objective is to spread anxiety, fear and panic by publishing violent videos like brutal executions and killings of people who are against the terrorist organizations. One of the most common methods of stopping these groups that spread propaganda when their accounts are discovered is to suspend (and/or completely block or terminate) these accounts.

However, this method requires human analysts who is responsible of manually analyze and scan a huge amount of published information on social media platforms. This paper proposed an automatic detection system that should be seen as a helpful ,complementary and accurate solution to detect the drastic content of such groups and present it to the analysts which he can take actions(to counter possible falsepositives). This paper presents a discussion of classification method to detect whether a tweet supports jihadistic groups or not. The text is classified into support , or against these groups.