Vijayakumar R
@rajalakshmi.org
Assistant Professor/CSE
Rajalakshmi Engineering College
Scopus Publications
Scholar Citations
Scholar h-index
Scholar i10-index
Scopus Publications
Vijay K, K. Jayashree, Vijayakumar R, and Babu Rajendiran
IEEE
Competition in gaming is one of the most widely audience watching team game in the ecosphere. Predicting the conclusion of a contest has developed so simple because to advancements in technology. We are utilising machine learning methods such as supervised learning to envisage the winner of a One Day International (ODI) cricket match. In order to train the models, we use each player's own career statistics as well as squad recitals such as flickering and speeding performances. The ability of players and many parameters, on the other hand, play an important role in determining the ultimate outcome of a match. As a result, we're employing supervised learning algorithms to forecast the match's outcome. It also aids the team's coaches in learning and analysing where the team's performance is lacking, so that the coaches may discover a solution to increase their team's strength. So, in this study, we use four different sorts of machine learning algorithms and compare them to each other to see which one is the most accurate and produces the greatest results.
Vijavakumar R, Ravikumar S, Vijay K, and Sivaranjani P
IEEE
In the event of an emergency, mobile text messaging systems are increasingly being used to disseminate analytical data. As a result, third-party companies claim to improve physical security by quickly delivering such communications to a variety of businesses, including colleges and universities. The primary goal of this study is to ensure user security during information sharing. The text messaging allows transmitting short, alpha numeric communication for a wide variety of applications. Public Alert System (PAS) presents two components mainly Administration Management and User Management. In PAS, there is a domain connected to intranet through which the user accounts can be created. Sending of mails to other user accounts by the authorized user is allowed and inbox can be managed. The system helps the users to view, store and delete the mails when necessary. Also, when the account is hacked, the hacker list along with the date, time of attack and the IP address is being stored in the user account. The user can view the hacker details by logging into appropriate account. In so doing, the system demonstrates that this infrastructure provides better security to the users since there is no hacking of mails.
Vijayakumar R, Vijay K, Sivaranjani P, and Priya V
IOS Press
The way of thinking of traffic observing for discovery of system assaults is predicated on a “gained information” viewpoint: current methods recognize either the notable assaults which they’re customized to alarm on, or those strange occasions that veer off from a known typical activity profile. These philosophies depend on an expert structure which gives the ideal data, either with respect to “marks” of the striking attacks or as anomaly free traffic datasets, adequately rich to make delegate profiles for commonplace movement traffic. The theory talks about the limitations of current information-based system to recognize organize assaults in an inexorably unpredictable and advancing Web, Described by ever-rising applications and an ever-expanding number of most recent system assaults. In an oppositely inverse viewpoint, we place the weight on the occasion of solo recognition strategies, fit for distinguishing obscure system assaults during a unique situation with none past information, neither on the attributes of the assault nor on the gauge traffic conduct. In view of the perception that an outsized portion of system assaults are contained during a little division of traffic flows, the proposition exhibits an approach to join basic bunching strategies to precisely distinguish and portray malignant flows. to bring up the practicality of such an information autonomous methodology, a solid multi-bunching-based location technique is created and assess its capacity to recognize and portray arrange assaults with none past information, utilizing bundle follows from two genuine operational systems. The methodology is acclimated identify and describe obscure vindictive flows, and spotlights on the identification and portrayal of ordinary and notable assaults, which encourages the translation of results. When contrasted with the predominant DDoS traceback techniques, the proposed system has assortment of favorable circumstances—it is memory no concentrated, proficiently adaptable, vigorous against parcel contamination, and free of assault traffic designs. The consequences of inside and out test and reenactment considers are introduced to exhibit the adequacy and effectiveness of the proposed strategy. It’s an uncommon test to traceback the wellspring of Circulated Disavowal of-Administration (DDoS) assaults inside the Web. In DDoS assaults, aggressors create a lot of solicitations to casualties through undermined PCs (zombies), with the point of keeping ordinary help or debasing from getting the norm of administrations. Because of this fundamental change, the proposed system conquers the acquired downsides of parcel stamping strategies, similar to weakness to bundle contaminations. The execution of the proposed strategy welcomes no changes on current steering programming. Moreover, this work builds up a hypothetical structure for assessing the insurance of IDS against mimicry assaults. It shows an approach to break the wellbeing of 1 distributed IDS with these strategies, and it tentatively affirms the capacity of various assaults by giving a worked model. The Project is intended by using Java 1.6 as face and MS SQL Server 2000 as backside. The IDE used is Net Beans 6.8.
Leena C. Sekhar and R. Vijayakumar
Springer Singapore
Natheeswari N, Sivaranjani P, Vijay K, and Vijayakumar R
IOS Press
Database Migration is a segment of a business application arrangement process wont to refresh changes in customer databases. This is frequently a significant procedure in application condition. Subsequently every association applies inventive information relocation strategies for refreshing their customer’s databases. These updates are typically more steady in nature than the intermittent updates performed under conventional discharge rehearses. Visit little updates make every movement less unsafe. Be that as it may, just in the event of tremendous updates, relocation process hinders prompting off base and inadequate changes. During this task, the HRAPP application framework’s database refreshes are moved utilizing another dependable and tedious strategy. We applied Continuous Integration and Continuous Deployment strategy for information relocation. Nonstop joining and arrangement pipeline technique might be a product improvement practice where designers consistently blend their code changes into a focal vault, after which computerized fabricates and tests are run. Here, arrangement code changes are consequently assembled, tried, and prepared for a discharge to creation. These movement documents are taken care of consequently through the sending pipeline, and changes to the databases are refreshed bolstered the relocation record. Right now, talk about how effectively database relocation are frequently executed with ceaseless joining and organization pipeline technique.
Vijay K, Vijayakumar R, Sivaranjani P, and Logeshwari R
IOS Press
This task depends on quality control in the vehicle business. It centers on the imprint and harms in new cars before producing to the client. This project presents the development of a system of recognition of defects and cosmetic imperfections in cars. This application gives a quick and strong robotized results. It likewise gives framework acknowledgement of scratches. In the as of now existing framework, the way toward distinguishing the scratches in car is finished by our mankind. Using the input frames, sections of the vehicles are entered for training, the last Fully-connected layer is altered so that it only has two exit categories: Sections with scratches and without scratches. This project is mainly developed to minimize manpower and maximize automation on quality department in automobile industry. It is a computer vision project. It includes task such as acquiring, processing, analyzing and understanding digital images and extraction of high dimensional data. An image processing algorithm is used in order to manipulate an image to achieve an aesthetic standard and to provide a translation between the human visual system and digital imaging services.
Sreesubha S, Babu R, Vijayakumar R, and Vijay K
IOS Press
A concealing picture is worked of various assortments of a comparative scene, one for each repulsive section. These dim level picture gives the gathering of light, each at the unequivocal awful part and at the circumstance of every pixel. Red, Green and Blue are three frightful gatherings. Each image is created of three diminish measurement pictures, three gatherings.We’ll conceal the substance in diminish pictures. Using Image improvement methodology, when we change the power estimations of basic tones by then the disguised substance is undeniable. From the preliminary outcomes we get 96% exactness
S. Rethishkumar and R. Vijayakumar
Springer Singapore
Leena C Sekhar*, , R Vijayakumar, M K Sabu, , and
Blue Eyes Intelligence Engineering and Sciences Engineering and Sciences Publication - BEIESP
The high dimensional dataset with irrelevant, redundant and noisy features has much influence on the performance of machine learning problems. In this work, an existing Ant Colony Optimization (ACO) based feature selection algorithm is modified by attaching a dimensionality reduction method as a data pre-processing step. This is achieved by introducing the concept of Centre of Gravity (CoG) of a set of points. After reducing the dimension, the ACO algorithm is used to generate the optimal subset of features. The performance of the proposed algorithm is evaluated using Artificial Neural Network (ANN) classifier. The performance comparison using various dataset shows that the proposed method outperforms the existing ACO based feature selection methods.
S P Srinivasan, J Anitha, and R Vijayakumar
IOP Publishing
The evolution of bio fuel supply chain has revolutionized the organization by restructuring the practices of the traditional management. A flexible distribution system is becoming the need of our society. The main focus of this paper is to integrate IoT technologies into a cultivation, extraction and management of Jatropha seed. It was noticed that major set-back of farmers due to poor supply chain integration. The various losses like information about the Jatropha seed availability, the location of esterification plants and distribution details are identified through this IoT. This enables the farmers to reorganize the land resources, yield estimation and distribution functions. The wastage and the scarcity of energy can be tackled by using the smart phone technologies. This paper is proposes a conceptual frame work on various losses involved in the supply chain of Jatropha seed.
Juby Mathew and R. Vijayakumar
IEEE
In various domains, big data play crucial and related processes because of the latest developments in the digital planet. Such irrepressible data growth has led to bring clustering algorithms to segment the data into small sets to perform associated processes with them. However, the challenge continues in dealing with large data, because most of the algorithms are compatible only with small data. However, the existing clustering algorithms either handle different data types with inefficiency in handling large data or handle large data with limitations in considering numeric attributes. Hence, parallel clustering has come into the picture to provide crucial contribution towards clustering large data. This insists the need of having scalable parallel clustering to solve the aforesaid problems. In this paper, we have developed a scalable parallel clustering algorithm called Possibilistic Fuzzy C-Means (PFCM) clustering to cluster large data. So, our ultimate aim is to design and develop an algorithm in parallel way by considering data. The parallel architecture includes, splitting the input data and clustering each set of data using PFCM. Then the genetic firefly algorithm applied to the merged cluster data, which will provide better clustering accuracy in merge data. The experimental analysis will be carried out to evaluate the feasibility of the scalable Possibilistic Fuzzy C-Means (PFCM) clustering approach. The experimental analysis showed that the proposed approach obtained upper head over existing method in terms of accuracy and time.
Juby Mathew and R Vijayakumar
IEEE
This paper mainly focuses on identifying the limitations of the K-Means algorithm and to propose the parallelization of the K-Means using Firefly based clustering method. The new parallel architecture can handle large number of clusters. Modified Firefly algorithm can be used to find initial optimal cluster centroid and then K-Means algorithm with optimized centroid can be used to refine them and improve clustering accuracy. The final convergence issue is also addressed and solved to a great extent. The design methodology is explained in the subsequent sections. Finally, modified algorithm is compared with Parallel K-Means. It is demonstrated with experiments and it has been found that the performance of modified algorithm is better than that of the existing algorithm. Four typical benchmark data sets from the UCI machine learning repository are used to demonstrate the results of the techniques.
Juby Mathew and R Vijayakumar
IEEE
This paper mainly focuses in identifying the limitations of the k means algorithm and to propose the parallelization of the k-means using firefly based clustering method. The new parallel architecture can handle large number of clusters. Firefly algorithm to find initial optimal cluster centroid and then k-means algorithm with optimized centroid to refined them and improve clustering accuracy. The final convergence issue is also addressed and solved to a great extent. Finally modified algorithm is compared with parallel k means is demonstrated with experiments and it has been found that the performance of modified algorithm is better than the existing algorithm. Four typical benchmark data sets from the UCI machine learning repository are used to demonstrate the results of the techniques. To achieve this we can use fork/join method in java programming. It is the most effective design method for achieve good parallel performance.
V K Kiran and R Vijayakumar
IEEE
NoSQL databases are suitable for network traffic loads experienced by today's websites. The important attributes that make them a favorable choice are inherent distributed nature, horizontal scalability and flexibility in data models. Hence, more data is now getting handled through NoSQL databases and sensible information extraction from these data stores is becoming a requirement. Information extraction may require sourcing data from multiple data sources, establishing relationship among them and querying across these data sources together. Ontology based semantic integration systems for Relational Database Management Systems (RDBMS) already exist that satisfies the above requirement. Many commercial systems are operational based on the above technique and tools and solutions for the above approach are very much mature. Hence, mapping the processing done in each stage of RDBMS based semantic integration systems to that of NoSQL systems can facilitate usage of existing tools and frameworks. The purpose of this work is to develop ontology based semantic integration system for a column-oriented NoSQL data store like HBase which is similar in architectural design to RDBMS.
R. Vijayakumar, K. Selvakumar, K. Kulothungan, and A. Kannan
IEEE
In wireless networks, spoofing attack is one of the most common and challenging attacks. Due to these attacks the overall network performance would be degraded. In this paper, a medoid based clustering approach has been proposed to detect a multiple spoofing attacks in wireless networks. In addition, a Enhanced Partitioning Around Medoid (EPAM) with average silhouette has been integrated with the clustering mechanism to detect a multiple spoofing attacks with a higher accuracy rate. Based on the proposed method, the received signal strength based clustering approach has been adopted for medoid clustering for detection of attacks. In order to prevent the multiple spoofing attacks, dynamic MAC address allocation scheme using MD5 hashing technique is implemented. The experimental results shows, the proposed method can detect spoofing attacks with high accuracy rate and prevent the attacks. Thus the overall network performance is improved with high accuracy rate.
Muhamed Ilyas and R. Vijayakumar
Springer Berlin Heidelberg
P. Muhamed Ilyas and R. Vijayakumar
ACM
Location dependant data are becoming very popular in mobile environments. To improve system performance and facilitates QoS several methods like location aware proxies, caching etc. are being used. Based on the classification of different query models and data, we propose an architecture with hierarchical service database for effective location based services to reduce the network cost with required quality of service (QoS). In our paper, a hierarchical database structure with collocated proxy is proposed to facilitate location based service delivery for both location dependant and independent data to the clients. A mobile user communicates with the proxy that is attached with the local (server) database of the cell, which routes the query to the appropriate level of the hierarchical database structure, based on the query type and location of the target object. A database update method has also been proposed based on the update frequency and access characteristics of data objects. We devise a general architecture and cost model for servicing location dependant and independent data, to reduce the network cost by maintaining quality of service and improve response time under changing user mobility and access patterns. Further when the mobile user engaged in multiple services, the per-service proxies can be employed to further improve the access time.
Vinu V. Das and R. Vijayakumar
IEEE
Ad-hoc wireless network is a collection of wireless mobile nodes forming a temporary network without any centralized administration. A typical paradigm of wireless networks, such as mobile ad hoc networks (MANETs), comprise a number of mobile nodes that can organize freely but need mutual cooperation for information exchanges. However, such networks also face several important challenges such as resource-limitation, changes in network topology due to the mobility of nodes; and moreover centralized network management is not possible. Considering all these challenges, its worth analyzing and finding the parameters, when a Secure System for Mobile Ad-hoc Network is designed. This paper illustrates three new design parameters and its sub parameters.