[1] Under Publication December 19, 2016 Toward A Comprehensive Terrorist Prediction in Social Network Ahmad F. Al Musawi Computer Science Department, College of Computer Science and Mathematics, Thi Qar University, Iraq ABSTRACT Social network analysis can play significant role in detecting human personality. Consequently, special personality characteristics can be analyzed to predict potential terrorism actions. These features could be conducted by using different social representations such as athletics, social and regional characteristics and so on. Data can be mined to result set of people with common features and categories. Based on that, people can be clustered according to their affiliation clustering on scale of dangerous and peaceful ones. Herein, a social network partitioning and clustering model is implemented to detect how close is a citizen to terrorism or a terrorist. KEYWORDS: Social Network Analysis, Terrorists Networks, Community Detection, Clustering. INTRODUCTION Predicting the possible activities and patterns of interactions among individuals or groups of individuals considered as an old and emerge field as well [1], [4], [5]. However, the results, roughly speaking, of such prediction can be extensively used to manage different fields of life as it will change the ways of implementing different life facilities such as curing, employing, leading, governing, trade marketing and much sophisticated approaches and techniques of strategic implementations. This prediction can play significantly in detecting suspects and terrorists in given society. Many academic and security institution and organization have focus their works on understanding the social connection among terrorists and their cover networks so that they determine the best strategy of defensing and attacking them [2], [6]. Predicting suspected people depends on the feature of each person and the social connectivity of that person. However, the features specification is very important. A terrorist is an ordinary person who have high weighted affiliation to specific features among others. Most researchers focus their attention on analyzing terrorists networks to searching for specific role terrorists among others[20],[21]. A person, from the informational perspective, is set of attributes that describe him/ her in any given time within his/ her environment. These attributes would have two different types: static and dynamic. Static attributes are these features that hardly be changed over time which can be his name, birthdate, eye color, hair color, figure print or any biometric feature, social security number and so on. Dynamic attributes are the personal features that simply be changed over time as a result of person interaction with his/ her environment, such as current position, health state, education, psychological mode, level of awareness, level of threading and so on. The categorical measurement specification of both the static and dynamic of human attributes may differ for many scholars and purposes. Computational social science deals with determination of these measurements. Many researchers conducted to measure different human affiliation to one activity/ behavior or another. These researches can be found in [3]. However, most scientists are focus on one-time, self- reported data on relationship [1]. The problem specification of this article is basically on how can we utilize the static information about a community for predicting terrorists within it, i.e. given social network of set of people and their attribute matrix, how could we categorize them based on their common behavioral attributes and connection? The social network depicts the relationships among people and the attribute matrix shows the different scaled measurement features of each person. Social network analysis [7] is the study of social interaction among people using any specified social relationship. Social network analysis SNA theories and algorithms have a wide range of contribution to other similar field such as disease modeling [8], [9], biological networks [10], [11], [12], [13], business management [14], [15], [16], [17], and more. A social network is the mathematical representation of interactions among people. The mathematical field of study used for presenting networks and its algorithms is graph theory. A graph , consist of a pair ሺ, ܧሻ,where  refers to set of vertices or nodes and ܧrefers to set of edges that presented as line connects one node with another. Individuals in SNA are presented as nodes (or vertices) and the relationships among them are presented as edges connecting them. The collection of relationships is presented as edges set, where each edge is a pair of individuals or their vertices. Better presentation of edges set is by using matrix representation. Let ܣbe an adjacency matrix of P×P values of integer number. ܣ ௜௝ ={ ͳ  ݐℎݎ  ݏ ݎݐ݋ݏℎ݌ ݓݐ ݌ݏݎ݋    Ͳ ݋ݐℎݓݎݏ ܣ ௜௝ would represent the relationship between the person in number  with the person in number . ܣis undirected network as the orient of relationship is from both direction. This network may represent any social relationship as friendship, marriage, work collaboration, joining same club and so on. It is very required to mention as much relationships as possible. The different collection of relationships can be implemented using extra arrays. LITERATURE REVIEW