Akhilesh Tiwari

On the Use of Brokerage Approach to Discover Influencing Nodes in Terrorist Networks

Social Network Analysis is a non-conventional Data Mining technique which analyzes social networks on web. The technique is used frequently for studying network behaviors using centrality measures viz. Degree, Betweenness, Closeness and Eigenvector. Hence has also led to the concept of Terrorist Network Mining which aims at detection of the terrorist group, studying the hierarchy they follow for the communication (using SNA) and then finally destabilizing of the network activities. The chapter focuses on an approach under SNA known as Brokerage which finds brokers who serve as the leading nodes in the network. Brokerage is expected to be beneficial in case of estimating the terrorist groups where different subgroups of terrorist organization coordinate to fulfill their awful deeds. The brokerage on whole estimates the influential roles as it would be done by individually calculating the centrality measures, with much more useful information aiding to amend terrorist network analysis.

A Novel Genetic Based Framework for the Detection and Destabilization of Influencing Nodes in Terrorist Network

The Social Network Analysis analyses social network on web. SNA has led the law enforcement agencies to study the behavior of terrorist networks for the identification of relationships that may exist between nodes. Recently, Terrorist Network Mining (special branch of SNA) has been in vogue in Data mining community because of its ability to identify key nodes present in the network. This paper proposes a new approach for Terrorist network mining. The proposed work is carried out in two phases. The first phase proposes Genetic based optimization mechanism. Proposed mechanism is suitable for effective optimization of large social network containing terrorist and non-terrorist nodes. During optimization process removal of non-terrorist nodes from the network has been performed and resultant represents the reduced graph containing only the set of potential nodes. The second phase proposes a weighted degree centrality measure (considers frequency of communication) for effectively neutralizing of the terrorist network.

Discovering Optimal Patterns for Forensic Pattern Warehouse

As the need of investigative information is increasing at an exponential rate, extraction of relevant patterns out of huge amount of forensic data becomes more complex. Forensic pattern mining is a technique that deals with mining of the forensic patterns from forensic pattern warehouse in support of forensic investigation and analysis of the causes of occurrence of an event. But, sometimes those patterns do not provide certain analytical results and also may contain some noisy information with them. An approach through which optimal patterns or reliable patterns are extracted from forensic pattern warehouse which strengthen the decisions-making process during investigations has been proposed in the paper.

A Proficient Video Compression Method Based on DWT & HV Partition Fractal Transform Function

Transform based function play vital role in video compression. In this paper used two transform functions for video compression one is discrete wavelet transform function and other is fractal transform function. The discrete wavelet transform function is very promising image compression technique. Instead of discrete wavelet transform fractal transform is fast image compression technique. Here both wavelet transform function and fractal transform function used for video compression. The H-V partition technique is used for fast processing of video data in terms of row and column. The process of compressing produce good PSNR value and the compression ratio instead of DWT transform function. The DWT transform function creates group of frames in terms of layer for the processing of lower and higher band data of process video. The major contribution of H-V partition technique in video compression due to represents of domain and range blocks for the processing of video components. The processing of transform generates the similarity property of range and the compression process is fast. The both methods DWT and H-V partition techniques simulated in MATLAB software and measure some standard parameter such as PSNR, Compression ratio, encoding time, and MSE.

Kohonen neural network model reference for nonlinear discrete time systems

In this work, an adaptive neural network like Kohonen neural network (KNN) model reference is used for tracking control of nonlinear system. Proposed adaptive Kohonen neural network (ADKNN) are used to minimize the error between output and target signal for nonlinear discrete-time systems. The ADKNN is a feed-forward neural network help for approximation of the nonlinearities in the industrial plant and main characteristic of the system is taken into account is disturbances in the system. Tracking error by the adaptive ADKNN based approximation system is an important characteristic for the design and analysis. It is shown in results that the preference of the error system is decisive to the solution of tracking control. Difference between ADKNN output and reference signal can be made arbitrarily small in the close neighbourhood of zero. The viability of the ADKNN is verified via simulation example of nonlinear system.

A novel way to classify passenger data using Naïve Bayes algorithm (A real time anti-terrorism approach)

Terrorist Data Mining basically means to encounter all the data of terrorism from the huge amount of data. In a more intricate way we all know that terrorist set their foot into any predominant place through railway station, bus stands or airport. Usually to communicate they use their cell phones and network. Now if these areas are well equipped with LAN or WAN, that is the Wi-Fi connections surely these terrorist would avail themselves. Then with the help of data used by terrorists their presence can be spotted easily and their information can be collected. This paper describes a novel work to counter the presence of terrorist at public place in a well-defined manner.

Tree-augmented naïve Bayes-based model for intrusion detection system

Despite enormous efforts for detecting unauthorised attempts to access a system or a network using an Intrusion Detection System (IDS), a major shortcoming still remains, which is the high False Positive (FP) rate, i.e. incorrect classification of the normal activities as abnormal (intrusion). It has been observed that the simple Bayes Net is one of the frequently used techniques for intrusion detection. Although satisfactory results have been obtained from the K2 algorithm incorporated in Bayes Net, the need for reducing the FP rate still arises. The present paper proposes a new model that serves as an alternative to Bayes Net with K2 algorithm, named TAN-based model for intrusion detection. This model has shown promising results with an advantage of more accurate detection of intrusions along with reduced FP rate.