......

..

IJCTA-Volume 7 Issue 2 / March-April 2016
S.No
Title/Author Name
Page No
1
Data storage Security in Single to Multi-Clouds Using TPA (Third Party Auditor)
-Vilas C. Rathod,Nitin Mishra
Abstract
Cloud Computing has been unreal as a result of the next-generation style of IT Enterprise. Cloud Computing could be a web process with great amount of resource. The user of the cloud will get the service thought network. It moves the applying software package and database bases to the centralized massive data centres, wherever the management of the data and services might not be absolutelytrustworthy.This distinctive paradigm brings concerning several new security challenges, that haven’t been well understood. Guaranteeing the security of cloud computing may be a major think about the cloud surroundings, as users typically store sensitive data with cloud storage providers however these suppliers is also untrusted. Handling with “single cloud” suppliers is expected to calm down in style customers as a results of risks of service availableness failure and additionally the danger of malicious insiders inside the only cloud. A movement towards “multi-clouds”, or in different words, “interclouds” or “cloud-of-clouds” has emerged recently. As we have a tendency to address 3 security factors that significantly have an effect on single clouds, specifically data integrity, data intrusion, and service availability.
So here we offer a framework to produce a secure cloud information which will guarantee to forestall security risks facing the cloud computing community. This framework can apply multi-clouds and therefore the secret sharing algorithmic program to cut back the chance data intrusion and therefore the loss of service accessibility within the cloud and guarantee data integrity victimization third party auditor (TPA). Third Party Auditor: An entity that has experience and capabilities that user don't have, is trusty to assess and expose risk of cloud storage services on behalf of the user upon request. Above all, we have a tendency to think about the task of permitting a third party auditor (TPA), on behalf of the cloud consumer, to verify the integrity of the dynamic information keep within the cloud.
224-229
PDF
2
Linked Unsupervised Based Advanced Feature Selection Framework with Artificial Bee Colony for Social Media Data
-
T.Pradeepa,B.Shanmugapriya
Abstract
The explosive usage of social media produces large amount of unlabeled and high-dimensional data. Feature selection has been proven to be effective in dealing with high-dimensional data for efficient learning and data mining. Unsupervised feature selection remains a challenging task due to the absence of label information based on which feature relevance is often assessed. Existing work investigates a novel problem of feature selection for social media data in an unsupervised scenario. Initially the work analyzes the differences between social media data and traditional attribute value data. Further it investigates how the relations are extracted from linked data which can be exploited to help in selection of relevant features using LUAFS. in LUAFS, social media networks have the availability of various link formations which leads to networks have the availability of various link formations which leads to networks with relationships of different strengths i.e., weak links and strong links that are often mixed together. Since strong links indicate strong correlations among instances, treating all links with an equal weight will increase the level of noise in the learned models and leads to degradation of learning performance. To overcome this issue, Artificial Bee Colony Algorithm has been introduced. A novel LUAFS-ABC has been proposed for linked data in social media to exploit linked information of selected features. To exploit the individual and group behaviors of linked instances two approaches: graph regularization and Social Dimension Regularization (SDR) have been developed. The experimental results of the data set from real-world social media websites shows that the proposed method can effectively exploit link information in comparison with the state-of-the–art unsupervised feature selection methods
INDEX TERMS: Unsupervised Feature Selection, linked data, social media, pseudo labels, social dimension regularization, Linked Unsupervised Feature Selection (LUFS), Artificial Bee Colony (ABC).
230-239
PDF
3
Rising future of Agile Software Development using Cloud Computing: A
study using Cloud Computing in different phases of an Agile method-SCRUM

-
Ritu Singhal, Sonia, Archana Singhal
Abstract
Due to the growing popularity of agile methods in the software engineering world nowadays companies do prefer and support agile methods over traditional methods. Agile methods are adaptive and provide software development at a faster pace with the flexibility of changes at any point of time. Also, during development of web applications, agile processes are gaining popularity as the prevailing conditions recognize that changes are inevitable. It is a lightweight and iterative approach most suitable with volatile customer requirements. However agile methodology needs a development platform todevelop software at a faster pace. This platform can be efficiently provided by cloud computing which accelerates the agile development. Thus, the present paper proposed a framework ESCAM which integrates the cloud services with an agile method SCRUM and describes how cloud activities can aggravate the SCRUM development activities.
Keywords: Cloud computing, Agile Software Development Methodology, Agile Methods, SCRUM
240-246
PDF
4
A Survey on Opinion Mining of Real Time Data using Big Data Analytics
-
Gargi Mishra,Shivani Varshney
Abstract
Social Networking sites provides tremendous impetus for giant information in mining people’s opinion. To observe individuals' feeling, tweets are labeled into positive, negative or nonpartisan pointers. This paper provides an effective mechanism to perform opinion mining from real time data from Twitter. First part of the research concerns crawling data from Twitter and string them into database. Second part of the research is mining the text to three major classification of positive negative or neutral. The classification done with various different approaches like naïve bayes, SVM (support vector machine) or ABC (artificial bee colony).
Index Terms – sentiment ; opinion mining; Web mining; web mining software; big data analytics ; social networking
247-252
PDF
5
Hop-Count Based Enhanced Co-operative Bait Detection Scheme using Prevention of Collaborative Blackhole Attacks in MANET
-
V.Abinaya,Dr. K.Santhi
Abstract
A MANET is a collection of mobile nodes connected through wireless networks. MANET can join and leave the network dynamically. However, MANET is particularly vulnerable due to its fundamental characteristics, such as dynamic topology, distributed co-operation, and constrained capability. One main challenge on designing these networks is their vulnerability to security attacks.In this paper the performance of Enhanced collaborative bait detection scheme(EnCBDS) using routing protocol AODV with Black hole attack detection have been analysed using NS2 considering various parameters such as average throughput, energy and end-to-end delay to evaluate its performance.
Keyword: Blackhole attacks, MANET, Enhanced Cooperative Bait Detection Scheme, Ad-Hoc On demand Distance Vector Protocol, hop count, malicious node.
253-260
PDF
6
Ring Spun Yarn Parameters Impact on Composite Yarn Quality Model
-
Hajer Souid,Mehdi Sahnoun,Morched Cheikhrouhou
Abstract
In the present study, we investigate the effect of fibers and construction parameters on the overall ring spun yarn quality. We develop for the case a global quality index for the ring spun yarn. This part of the study was achieved by using Derringer and Suich desirability functions. Then, we have applied an artificial neural networks model to these parameters. We base our research work on a prediction model to represent global ring spun yarn quality. This model is developed by optimizing fibers parameters, yarn count and twist. Finally, we tried to search the impact of all these parameters on the neural networks model while aligning a confidence interval.
Keywords: ring spun yarn; quality; desirability functions; neural networks; confidence interval.
261-267
PDF
7
Designing of Network Model to Classify Wireless Sensor Networks
-
Siva Balaji Yadav.C, R.Seshadri
Abstract
Distributed Denial of Service (DDoS) attacks are defined as attacks launched from multiple ends of a wireless sensor network towards a set of legitimate sensor nodes, with the intent of exhausting their limited energy resources. These attacks can significantly affect the performance of the network, and eventually lead to complete compromise of all sensor nodes of the network. The consequences of such an attack, if left undetected, can be catastrophic to the operations of the entire network. The distributed nature of the proposed algorithm ensures that most steps of the attack detection process are performed within the sensor network, without the need to communicate on a frequent basis with centralised network base stations. Several optimization criteria, such as frequency of convergence of the detection scheme, and selection of specific detector and decision-making nodes, are addressed as part of the detection schemes to reduce the overhead incurred on the sensor resources. We also perform an evaluation of the scheme through simulation experiments, to test the effectiveness of our approach. In addition, the quantitative results acquired from the experiments are benchmarked with corresponding results acquired from a centralised Self-Organising Map-based attack detection scheme. Through the result comparisons, we prove the significance of distributed pattern recognition in such networks, for detecting distributed denial of service attacks in a timely and energy-efficient manner.
Key words: Distributed Denial of Service, Algorithm, sensor network
268-273
PDF
8
Naming Disambiguation Based on Approximate String Matching for Co-Authorship Networks
-
Dr.V.Akila,Dr.V.Govindasamy,R. Kowsalya
Abstract
Co-authorship network is a network that models the co authorship of scientific publication in a network. Naming Disambiguation is an important aspect of Co-authorship Network. To finding an unique author in a co authorship network is a challenging one. In co authorship network multiple persons have the same name, name abbreviation, name misspelling etc. Further, human error leads to considering multiple persons under a single reference. Such mistakes affect the performance of finding an unique author. Author name disambiguation is to assign a unique identifier to the same name. A naming disambiguation model based on approximate string matching algorithms Jaro wrinkler and Levenstien similarity is proposed in the paper. The proposed System Naming Disambiguation pertains to assigning an unique ID to each unique author.
Keywords : Disambiguation; Co-Authorship Network; Approximate String Matching
274-278
PDF
9
Performance Analysis of Apriori and FP-Growth Algorithms (Association Rule Mining)
-
Alhassan Bala,Mansur Zakariyya Shuaibu,Zaharaddeen Karami Lawal,Rufa’i Yusuf Zakari
Abstract
Association rule mining has become popular among marketers and organizations. In fact, an example of association rule mining is referred to as market basket analysis. The task is to find which items are frequently purchased together. This knowledge can be used by professionals to plan where to place items that are frequently bought together closely to each other, thus helping to improve the sales. It involves the relationships between items in a data set. Association rule mining finds out item sets which has minimum support and are represented in a relatively high number of transactions. These transactions are simply known as frequent item sets. The algorithms that use association rules are divided into two stages, first is to find the frequent sets and the second is to use these frequent sets to generate the association rules. In this paper we used Weka to compare two algorithms (Apriori and FP-growth) based on execution time and database scan parameters used are; number of instances, confidence and support levels it is categorically clear that FP-Growth algorithm is better than apriori algorithm.
Keywords: - Association, Instances, Support, Confidence, Weka
279-293
PDF
10
Robots as Teachers: The Beginning of an Era
-
Dachapally Prudhvi Raj
Abstract
The rapid development in applications of Artificial Intelligence enabled the world to use those concepts in a myriad of disciplines. With the increase in Machine Learning, and Large-Scale Data Processing, scientists and researchers around the globe are now getting closer to create machines that are as intelligent as humans. The field of AI being very broad encompassed itself into industries such as health-care, investment predictions, business decisions etc. Though it has made it ways into many fields of study, the development of these technologies has not been up to the mark in the domain of education. The applications of most of the modern machine learning techniques are not used in the field of education. They are to a minimal extent such as marks predictions, intervention assessments, quiz evaluation etc., individually but not cohesively. This work showcases, that considering the astounding works in the subjects of Robotics, Artificial Intelligence, and Machine Learning, robots can be created which can act as assistants for the faculties, or can even acquire tenured positions in the universities. When the complete process of creating an artificial embodiment of a human teacher becomes feasible and functional, these entities should be able to replace the human kind in schools across the map.
294-299
PDF
11
An Empirical Study of Supervised Learning Techniques on Multispectral
Dataset

-
Arun Sharma,Dr. Vidushi Sharma
Abstract
In recent years, machine learning played a vital role in the field of artificial intelligence. Techniques like classification and regression has wide application in various field especially in data mining. In remote sensing, classification of objects has shown great significance such as in disaster management and defense purposes. In this paper, we studied supervised learning techniques on multiclass dataset extracted from multispectral satellite image. Investigation while simulating certain parameters of classifiers such as ANN, SVM, K-nn etc. gave some promising results. Analysis and comparison depends on performance measure like accuracy precision, recall and F-measure. Amongst these techniques, K-nearest neighbor has given the best result with 90.3% accuracy for multispectral satellite image dataset.
Keywords: Classification, Regression, ANN, SVM, Knn (K-nearest neighbor).
300-304
PDF
12
Objective Cost Based Replicas Relocation Services in Data Intensive Grid Environment
-
P.SunilGavaskar
Abstract
In order to manage large scale of grid, researchers always use to perform dynamic access; On behalf of that the replica management becomes a key idea that enhances the data access performance. The computing resources and data access resource are widely used in replica management for performance measurement. Such kind of data access resources allocation must be performed based on Time, Cost, Priority, Profit, Optimization and Quality of service methods. In our proposed strategies we can adapt a functionality called objective cost that is based on ‘updates of replicas status’ that happens data access environment, thusly replicas access is considered into an account to prototype the part of data access and its network service performance in grid environment.
In endorsement of replication issues there is presence of an additional overhead while performing updates and is due where data availability mostly applicable, and thus the status failures of replicas will take place. To compensate the changes that are occur on replica locations and its movement measurement, In this paper the author is considered objective cost factor for relocation mechanism, changes announcement at each stage while status updates along with replicas and its data access parameters that are involved in replica location and its status changes.
In general we consider that the placement of replica in the parent node: In common such kind of placement of replicas generates the maximum request on parent node. We can able to resolve this problem by measuring search time, distributed network loads between nodes. Therefore the access load can be defined as the cumulative total number of requests that is responded by a node in grid, where the requests are received from all of its children.
Our proposed algorithm in this paper can leads to produce better results at the time of additional overhead occurred while performing updates and its failures in the status of replicas.
IndexTerms— Data grid, Object Replicas, Replication, Resource Access Cost, Network Services, Replication Methods. Data Intensive Grid Environment
305-311
PDF
13
Designing a Smart Post-Surgery Recovery Ward through Service-oriented Context-awareness
-
Ashok Immanuel V,Pethuru Raj
Abstract
A variety of our social, personal, and professional environments are enabled to be smart through a host of technologies. The embedded smartness automates several manual tasks, accelerates and augments things in order to ensure enhanced care, choice, convenience, connectivity and comfort for people. In this paper, we have described how service composition techniques come handy in dynamically extracting context information precisely in order to establish and sustain a context-aware hospital environment, which can facilitate doing things intelligently. We have explained the use case in detail, articulated the major sensor and device services, implemented them in the RESTful format, and demonstrated how various services in runtime identified, orchestrated and leveraged to precisely understand the brewing situation and fulfil various healthcare-related activities automatically
312-316
PDF
14
Ontological Approach to Modeling a Learner in Adaptive learning Environment
-
Zakaria LAHBI,Mohamed SABBANE
Abstract
Today, with the fast development of the "world wide web (www)" and E-learning technology, personalization and adaptation of content becomes more and more a necessity. This personalization improves not only access to information but also the quality of teaching. Moreover, support for learners in their learning path requires a suitable adaptation of content.
In this work, we aim to model the various elements and attributes related to an ALE (Adaptive Learning Environment) especially for learners. In this perspective, we build ontology to represent a learner profile and relationships between the different attributes that an ALE may contain.
Keywords—component; E-learning; Ontology; ALE; Adaptation; Personalization; learner profile
317-321
PDF
15
Survey of Traffic Redundancy Elimination Systems for Cloud Computing Environment to Reduce Cloud Bandwidth and Costs
-
Ashwini Ghorpade,S.R.Idate
Abstract
This paper proposes a survey on various approaches of traffic redundancy elimination (TRE) system, which is useful in cloud computing environment. The traffic redundancy elimination system is based on predictive Acknowledgement (PACK) traffic redundancy system. Cloud computing is known as a pay-as-you-go service, with which cloud users uses various techniques for elimination of redundancy for cost and bandwidth. Transmission cost has significant role to minimize the huge cloud cost occurred during transmission. Numerous of existing server specific TRE approach cannot handle the traffic effectively. Several TRE practices are unsuitable for the cloud environment because the processing costs are high. This survey depicts a primary survey of the new traffic redundancy system known as novel-TRE also known TRE using PACK. The salient features of PACK technique is to detect the redundancy at client side, replications are come out in chains, received chunks are matched with the earlier times received chunk chain and transfer it to the server for prediction of future data. The benefit of this system is there is no requirement of server to maintain client state continuously.
The abstract is to be in fully-justified italicized text, at the top of the left-hand column as it is here, below the author information. Use the word “Abstract” as the title, in 12-point Times, boldface type, centered relative to the column, initially capitalized. The abstract is to be in 10-point, single-spaced type, and up to 150 words in length. Leave two blank lines after the abstract, then begin the main text.
322-327
PDF
16
An Integrated Isometric and Suppression Based Privacy Preservation Clustering on Multi Dimensional Data
-
A.Malaisamy,Dr.G.M.Kadhar Nawaz
Abstract
Currently most of the applications utilizing data mining techniques work heavily on private and sensitive information. Anonymizing the data and preserving the privacy of the dataset was proposed in most of the literature by preserving sensitive information for data mining. Most previous work on privacy preserving framework focuses on anonymizing data to meet privacy requirements or through oblivious transfer. However, the preservation of confidential attributes at the cost of privacy and anonymity poses serious problems. In this work to improve the preservation of confidential attributes in clustering analysis and to improve the anyonymity level an integrated framework called Isometric and Suppression-based Privacy Preservation Clustering (IS-PPC) on multi dimensional data is introduced. Initially, an Isometric-based Additive Noise Perturbation Transformation method is designed to preserve the intrinsic attribute values by transforming the confidential attributes at a time. The data matrix used in isometric-based transformation is subjected to clustering contains both the confidential and non-confidential attributes to protect individual data values before clustering. Next, suppression model to anymous the database based on the quasi identifier is designed aiming at improving the anonymity level. The effectiveness of our framework is demonstrated by a thorough evaluation and comparison over Adult dataset extracted from UCI repository. The performance improvement thus achieved makes Isometric and Suppression-based Privacy Preservation Clustering on multi dimensional data superior to other state-of-the-art woks. Experiment is conducted on factors such as execution time to preserve privacy, anonymity level, privacy preservation rate and data mining accuracy. Experiment results show that the proposed framework achieves better performance in improving the data mining accuracy by 10.05% and improves the privacy preservation rate on multi dimensional data by 13.83% compared to the state-of-the-art works.
Keywords: Anonymizing, Suppression-based, Privacy Preservation, Isometric-based, Additive Noise Perturbation, Transformation
328-336
PDF
17
Efficient and Secured Approach for Inter-firewall Optimization and Communication
-
Trupti Inamdar,Prof.R M Goudar
Abstract
Internet has been widely using the firewalls for securing private networks. To improve the network performance optimization of firewall policies is crucial. Previous work on optimization of firewall attentions on either intra-firewall or inter-firewall optimization within single administrative domain where the privacy of firewall policy is not taken into consideration. The challenge here is that firewall policies cannot be disclosed across domains because a firewall policy holds secret data and also security holes, which can be misused by attackers. Proposed work involves first privacy ensuring optimization protocol for cross domain firewalls. Precisely for any two adjacent firewalls from two different administrative domains, rules in each firewall are identified. The process of optimization includes co-operative computation between the two or more firewalls without any side disclosing its policy to the other for security purpose. The communication cost is reduced to less than a few hundred kilobytes. This protocol does not acquire additional packet processing overhead over the internet and the offline processing time is also less.
Keywords: Interfirewall optimization, security, redundancy
337-341
PDF
     
IJCTA © Copyrights 2015| All Rights Reserved.

This work is licensed under a Creative Commons Attribution 2.5 India License.