期刊名称：International Journal of Computer Science & Technology
印刷版ISSN：2229-4333
电子版ISSN：0976-8491
出版年度：2017
卷号：8
期号：1
页码：71-73
语种：English
出版社：Ayushmaan Technologies
摘要：In current scenario internet become as a de-facto medium of the communication to others. Internet provides an easy and efficient way of communication. There is a huge amount of content flooded over the internet which contains rational and extremist data. That data spread into various social media websites and social media communities to recruit people for ration and violent extremist activities. Contents like white hated music creates hated perception in the people and approach them to work for any terrorist organization or promote nationalism. A review over the techniques which are used to detect and forecast such activities is presented in this paper. Topic models like LDA (Latent Dirichlet Allocation) and CTM (Correlated Topic Model) etc. are used to predict such activities and ARIMA or some other models are used to forecast such activities. A new topic model called DCNT (Doubly Correlated Nonparametric Topic Model) and a multiple word matching or n-gram matching technique is proposed to provide better performance for detecting and predicting such violent extremist activities.
