Informatics Publishing Limited

A. Kumaravel ¹, R. Udayakumar ²

Affiliations
1 School of Computing, Bharath University, Selaiyur, Chennai-600073, IN
2 School of Computing Sciences, Bharath University, Selaiyur, Chennai-600073, IN

Abstract

Web mining is applied to reflect the importance of webpages and to predict the web domain visits of various users. An IF-inference system is developed for this purpose. This paper presents basic notions of web mining with fuzzy inference systems based on the Takagi-Sugeno fuzzy model. At first, it scans the sequences of visits in database once, finds all the weighted frequent items, and produces a sparse matrix for that sequence item set from which basic statistics like moving averages and exponentially smoothing averages are calculated. Then we construct a IF-inference system using intuitionistic approach for predicting the web domain visits. We obtain the results of such system.

Keywords

Web Mining, If-inference System, Web Domain Visits, Fuzzy Logic, Prediction, Sparse Matrix, Moving Averages, Exponential Smoothing

Full Text

   

A. Kumaravel ¹, K. Rangarajan ²

Affiliations
1 School of Computing, Bharath University, Chennai-600073
2 Dept of Computer Applications Bharath University, Chennai-600073

Abstract

A variety of interesting problems arise in the study of finite automata that move about in a two dimensional space. The model proposed by Muller [4] is used here to construct new automaton which can explore any labyrinth and escape through the moving or dynamic obstacles inside over the grid. The earlier results were shown for static obstacles distributed over integer grid and the automaton in this case was constructed to interact on the rectangular grid location endowed with four neighborhood directional states. In this paper we allow obstacles moving in discrete steps and verify that the finite automaton with just five printing symbols can escape or find the exit.

Keywords

Labyrinths, Finite State Automaton, Printing Mouse, Free Way of Escape

Full Text

   

A. Kumaravel ¹, Oinam Nickson Meetei ¹

Affiliations
1 Department of Computer Science & Engineering Bharath University, Chennai

Abstract

One of the core issues for robust cryptographic systems is to deal with the rate of diffusion and distribution of keys. We try to strengthen this aspect by increasing the length of block ciphers. We present a new encryption and decryption algorithm for block cipher that supports 27 bit block size. All components in our system are chosen to be based on cellular automata so as to achieve higher parallelism and to simplify the in hardware and software implementation for applications with high degree of security. The main objective of this paper is to increase the complexity by novel schemes of mixing (reversible cellular automata) RCA and (non-uniform reversible cellular automata) NRCA. We apply set of different bit permutation methods for this purpose .This paper establishes the proof for existence of yet another approach for a high quality pseudorandom bit sequences generated by non uniform cellular automata.

Keywords

Cryptographic System, NRCA-non Reversible CA, Pseudorandom Number, RCA-reversible Cellular Automata

Full Text

   

A. Kumaravel ¹, R. Pradeepa ²

Affiliations
1 Department of Computer Science and Engineering, Bharath University, Selaiyur, Chennai-600073, IN
2 Department of Computer Science and Engineering, Bharath University, Selaiyur, Chennai - 600073

Abstract

Subcellular localization is a well-designed representation of proteins. We need a fully automatic and reliable prediction system for protein subcellular localization, especially for the analysis of large-scale of yeast microarray data. In this paper we consider the dataset with multi classes and propose the classification for each location of protein subcellular in a separate layer. In this work, a multi-classification approach for subcellular localization is designed and developed to achieve high efficiency and improve the prediction and classification accuracy. The rule based Ripper method has been found to predict the subcellular localization of proteins from their protein microarray data, compared to other classifiers.

Keywords

Data Mining, Microarray, Classification, Layered Approach, Protein Subcellular Localization

Full Text

   

R. Udayakumar ¹, A. Kumaravel ², K. Rangarajan ²

Affiliations
1 Department of Information Technology, Bharath University, Chennai-600073
2 School of Computing, Bharath University, Chennai-73, IN

Abstract

Implementing a distributed process using object oriented programming is challenging especially for clusters of nodes ensuring the availability. Java happens to be well suited for writing object oriented programs for applications which needs modularity and high degree of cohesion [2]. In this paper our objective is to propose appropriate programming paradigms for loosely coupled distributed and object-oriented systems with a specific focus on parallel programming and client-server object computing. We select a prototype tool made based on the research outcome as found in [1] for this purpose. This research prototype tool 'Vishwa' provides a very good starting point for graduate students, engineers and scientists in academia, industry and government to present their tasks using the aspects of parallel and distributed computing. The learning curve is shorter in the domain of grid due to the simplicity in deploying and processing with this tool. We also demonstrate with a case study for bringing out the easiness in splitting the given job into subtasks and executed in this type of grid environment.

Keywords

Grid Computing, Distributed Computing, Object Oriented Programming, Speedup Ratio, Middleware

Full Text

   

S. Meher Taj ¹, A. Kumaravel ²

Affiliations
1 Department of Mathematics, AMET University, Chennai - 603112, Tamil Nadu, IN
2 School of Computing, Bharath University, Chennai - 600073, Tamil Nadu, IN

Abstract

The aim of this study is based on Survey on fuzzy Petri nets for Classification. Petri Nets (PN) is excellent networks which have great characteristics of combining a well defined mathematical theory with a graphical representation of the dynamic behavior of systems. But, with the growth in the difficulty of modern industrial, and communication systems, PN found themselves inadequate to address the problems of vagueness, and imprecision in data. This gave rise to combination of Fuzzy logic with Petri nets and a new tool emerged with the name of Fuzzy Petri Nets (FPN). Much research has been done for FPN and a number of their applications have been expected, but their basic types and structure is still ambiguous. Result of this research, an effort is made to categorize the applications of FPN for classification according to their structure and algorithms. We identify the different types of Petri nets will improve future research on Petri nets. Hence in this study, due to these limitations we focus on establishing the FPN in the light of their classifications has been done.

Keywords

Discrete Event Systems, Fuzzy Logic, Fuzzy Petri Nets, Petri Nets

Full Text

   

A. Vinu ¹, A. Kumaravel ²

Affiliations
1 Department of Information Technology, Jerusalem College of Engineering, Chennai - 600100, Tamil Nadu, IN
2 Department of Information Technology, Bharath University, Chennai - 600073, Tamil Nadu, IN

Abstract

Instant Messengers become an important means of communication. Millions of people. regardless of age, nationality, gender and computer skills, spend a lot of time using them every day. Thus, Social networks already occupied the place of traditional messaging systems of the past. More and more communications are migrating from public chat rooms and private messengers into online Social Networking Sites (SNSs). As cybercrimes mushroom in recent years, more and more digital crime investigations have strong relations to these SNSs. Communications extracted from SNSs can be extremely valuable and useful to all kinds of investigators including forensic investigators. We now apply the spotlight on distinct strings specific to each SNSs and volatile memory analysis and display of any case detail as a key to a successful digital investigation through rapid & improved retrieval performance.

Keywords

Digital Evidence, Distinct String Analysis, Forensic Investigators, Instant Messengers, Volatile Memory Analysis

Full Text

   

X. Anita ¹, A. Kumaravel ²

Affiliations
1 Department of Information Technology,Jerusalem College of Engineering, Chennai - 600100, Tamil Nadu, IN
2 Department of Information Technology, Bharath University, Chennai - 600073, Tamil Nadu, IN

Abstract

WSNs comprises of densely deployed tiny sensor devices deployed in sensitive applications. The security systems designed does not consider the constraints of sensor nodes. Trust of the network depends on packet forwarding behavior of the neighbor. The communication overhead depends on the number of packets transmitted for trust computation. The memory consumption depends on the size of the trust tables used for trust computation. In this paper, we have presented the communication overhead and memory consumption of the existing trust based routing schemes.

Full Text

   

Pallab Dutta ¹, A. Kumaravel ²

Affiliations
1 Department of CSE, Bharath University, Chennai - 600073, Tamil Nadu, IN
2 School of Computing, Bharath University, Chennai - 600073, Tamil Nadu, IN

Abstract

Recommender Systems has been a research hotspot in recent times as an efficient information filtering tool, to filter out useful required information from ever expanding web. The characteristics of social networks play a very important role towards behavioralmodeling of a trust network based Recommender System (RS). Similar to real world, in a social network also, it is important to objectively identify a member with high reputation who is heavily trusted by many members and hence his suggestions and inputs are most trust worthy for the whole community; that is, objective identification of leader is extremely important. In this paper, we propose a method to objectively identify leaders in a social network, we introduce new terms: Leadership score, prominence trust, peer inclusive factor, trust spread factor, trust maturity factor and trust penetration factors to portray a member more appropriately in a social network. We have calculated the leadership score using the real world dataset. Leadership score is taken as a linear additive function of prominence trust, engagement trust and peer inclusive factors with different weightage values. In our experiment, we have considered that leaders not only are followed by others, they are also socially engaged from their end to others in the social network. Prominence trust is a detailed characterization to get more accurate value of trust. Along with overall trust score, it is imperative to analyze more attributes to derive a more objective interpretation of the same. These attributes take care of trust score/reputation over period of time, number of trusting members, peer level interactions and also absolute score. To validate our leadership model in social network, we remove top 5%, 10% and top 20% of the members with high ‘leadership score’ and findout the reduction in the overall interactions. As further improvements to this work, a social network to be built with users and items and collect the data using positive and negative responses; use multiprogramming concepts to determine the optimum function for leadership score. This will aid towards further analysis and fine tuning of the models.

Keywords

Data Mining, Recommender System, Social Network, Trust, Web Mining

Full Text

   

A. Kumaravel ¹

Affiliations
1 Department of Computer Science, Bharath University, Chennai - 600073, Tamil Nadu, IN

Abstract

We address the problem of constructing an efficient route algorithm for Mobile Ad-hoc Network setting up QoS routes based on route discovery mechanism of AODV protocol and cellular automata. A Mobile Ad-hoc Network (MANET) consists of mobile nodes which are free to move arbitrarily due to its dynamic nature, adaptive and infrastructure less in nature. It is very tough task to have a snap of a network topology in this dynamic nature. Many research works have done to find a shortest path in MANET, it may give a better solution for finding an optimal path but difficult to augment with QoS constraints and its energy utilization. Here we have proposed an algorithm based on hexagonal cellular automata model for giving an efficient algorithm. This algorithm applies ADOV protocol and Lee's algorithm for constructing a shortest route and it considers energy and delay/bandwidth constraints as its basic metrics. These two metrics give a basic criterion for selecting a path between source and destination nodes through hexagonal cellular automata. Finally it is modeled with a matrix to implement it.

Keywords

ADOV, Cellular Automata, Hexagonal Grids, Lee Algorithm, Mobile Ad-hoc Networks, Quality of Service.

Full Text

   

G. Michael ¹, A. Kumaravel ², A. Chandrasekar ³

Affiliations
1 Department of Computer Science and Engineering, Bharath University, Chennai, IN
2 School of Computing, Bharath University, Chennai, IN
3 St.Joseph College of Engineering, Chennai, IN

Abstract

We address the problem of malicious node detection in a network based on the characteristics in the behavior of the network. This issue brings out a challenging set of research papers in the recent contributing a critical component to secure the network. This type of work evolves with many changes in the solution strategies. In this work, we propose carefully the learning models with cautious selection of attributes, selection of parameter thresholds and number of iterations. In this research, appropriate approach to evaluate the performance of a set of meta classifier algorithms (Ad Boost, Attribute selected classifier, Bagging, Classification via Regression, Filtered classifier, logit Boost, multiclass classifier). The ratio between training and testing data is made such way that compatibility of data patterns in both the sets are same. Hence we consider a set of supervised machine learning schemes with meta classifiers were applied on the selected dataset to predict the attack risk of the network environment. The trained models were then used for predicting the risk of the attacks in a web server environment or by any network administrator or any Security Experts. The Prediction Accuracy of the Classifiers was evaluated using 10-fold Cross Validation and the results have been compared to obtain the accuracy.

Keywords

Meta Classifier, Data Mining, Decision Trees, Decision Rules, Malicious Detection, KDD Dataset, Machine Learning, Network Security.

Full Text

   

C. Raghavendra ¹, A. Kumaravel ¹, S. Sivasuramanyan ²

Affiliations
1 Depatment of CSE, Bharath University, Chennai – 600073, Tamil Nadu, IN
2 Depatment of CSE, Dhanalakshmi engineering College, Chennai – 601301, Tamil Nadu, IN

Abstract

Objective: Iris recognition is one of the emerging areas as the demand for security in social and personal areas is increasing day by day. The most challenging step in the process of iris recognition is accurate iris localization. Methods/Statistical Analysis: In this paper, we propose an iris recognition method in light of Teaching Learning Based Optimisation (ITLBO) to choose the ideal components subset. The iris information, for the most part, contains a tremendous number of textural elements and a similarly modest number of tests per subject, which make the accurate iris pattern classification challenging. Findings: Feature selection scheme is used to identify the most important and irrelevant features from extracted features set of a relatively high dimension based on some selection criterions. It is not generally handy to gather an extensive number of tests because of some security issues. In this paper, we propose ITLBO to enhance the feature subset determination by consolidating important results from different component choice strategies. Application/Improvements: The principle target of ITLBO is to accomplish an adjust the Recognition Rate (RR), the False Acceptance Rate (FAR), the False Reject Rate (FRR) and they chose feature subset measure. The proposed method is computationally successful with the RR of 97.97 % on the CASIA iris datasets.

Keywords

Biometrics, CASIA, Feature Subset, ITLBO, Log Gabor Filter

Full Text