International Journal of Recent Technology and Engineering (IJRTE) ISSN: 2277-3878, Volume-8 Issue-4, November 2019 1467 Published By: Blue Eyes Intelligence Engineering & Sciences Publication Retrieval Number: D7590118419/2019©BEIESP DOI:10.35940/ijrte.D7590.118419 Data Aggregation and Terror Group Prediction using Machine Learning Algorithms S P Maniraj, Deep Chaudhary, Vankayala Hari Deep, Vishesh Pratap Singh Abstract : This paper is about to introduce a proposed system that examines growth or decay of the terrorist groups by the time, active locations, types of attack they carry out, motive targets, Weapon mastery and availability and many parameters to analyze the patterns and hidden structures in their activity and to predict the occasion and type of their future attack. We have done a detailed analysis of data we get from different sources and we also performed different classification algorithms on the available data to find the chances of probable attack on different regions.Based on results finding which of the algorithms works with highest accuracy. Key Words: Analysis, Classification, prediction, terrorism, Data aggregation, Terror Group. I. INTRODUCTION Over the past many years the world has been a witness for the remarkable number of terrorist events because of the terrorist event, the main victim is people the terrorist event which is happening in the world is not in the random order each terrorist event is interlinked with other in the world. We used to follow a particular pattern that initiates the terrorist activity; our task is to predict the event before they initiate it. To obtain the real- time data of the past terrorist events in the world and implement the clustering and data aggregation with the algorithms like logistic regression, SVM, K-NN we analyze clusters related to four combinations terrorism - event terrorism – target , terrorism target-terrorism agencies , terrorism agencies- terrorism attack type ,terrorism attack type-terrorism method, terrorism method-victim location. Through dataset, intuition is to annually analyzing the number of content of the cluster, effectively from 1980 to till date. Based on the clustering and data classification mechanism work is to predict the terrorist group responsible based time duration, event place, agencies, target, and attack type from the input and predicting terrorist group as output.According to the Global Terrorism Data report there is certain criterion to be chalked out to decide the objective and find the solution. II. RELATED WORKS 1.The paper published by Dr.Tariq Mahmood and Mr.Khadija Rohail has suggested worked on applying the use of data minining methods on terror activities in Pakistan. The study was focused on the cluster analysis which is used to group all raw data together in form of clusters as they possess common property, so it was found out to be ClOPE Revised Manuscript Received on November 15, 2019 S P Maniraj, Assistant Professor (Senior Grade), Department of Computer Science and Engineering, SRM Institute of Science and Technology, Chennai (Tamil Nadu), India. Deep Chaudhary, Student B.Tech, Department of Computer Science and Engineering, SRM Institute of Science and Technology, Chennai (Tamil Nadu), India. Vankayala Hari Deep, Student B.Tech, Department of Computer Science and Engineering, SRM Institute of Science and Technology, Chennai (Tamil Nadu), India. Vishesh Pratap Singh, Student B.Tech, Department of Computer Science and Engineering, SRM Institute of Science and Technology, Chennai (Tamil Nadu), India. as a proficient algorithm to analyze clusters on the raw data to analyze the terror groups or active organizations of any area as their study was specifically based on 4 provinces of Pakistan it provides the clustered data in form of target-type, target party, locations and several other factors that distinguishes them. 2. The paper on Effectiveness of Terrorism policies by Chaomin Lou and yang li suggests that their work is used to make the assumptions based on theory evaluation and then it test the effectiveness of US counter terrorism policies by implementing a time series interrupted model. This model is used to improve as it categorizes the policies based on their proactiveness it simplifies a time series equation that give results on basis of some mannequin parameters that are used to rate the policies in three categories as proactive, policy that works on the root of terrorism and also insignificant policies it worked well as time series functionality improves with the increase in the number of training data it was a commendable works as it helps to determine active policies and their model was tested on insurance policies too which gave them some excellent results. 3. Most related work was to form a predictive modeling model that gives warnings about the future assaults on Pakistan by two scholars Hina Muhammad Ismail and Abdullah Kazi their journal suggests the use of classification modeling. The study was to analyze the previous incident records which were available in GTD database. This study was focused specifically to Pakistan it takes certain common parameters which helps in identifying the threats to type of targets and locations that helps security agencies to work and make decision based on collective feed by the previous records to aid them for keeping security measures. III. PROCESSING A.ARCHITECTURE MODEL Based upon the proposed system it is essential to depict the significance of each step for elaborating the method or techniques used for drawing out inference based on the given dataset Model is, collecting data is used to capture records of past events to analyze data and find the recurring patterns to draw out inference that which data is useful and which data is of no use. Data is collected from GTD (Global Terrorism Database). B.DATA COLLECTION In this phase it is necessary that the data we collect is gathered from relevant sources. The quantity and quality of data tells how accurate the On terrorism which is gathered from relevant sources. Architecture model represents the