Global Journal of Computer Science and Technology (GJCST)
Not a member yet
1830 research outputs found
Sort by
A Survey on Clustering Techniques for Multi-Valued Data Sets
Get PDFThe complexity of the attributes in some particular domain is high when compare to the standard domain, the reason for this is its internal variation and the structure .their representation needs more complex data called multi-valued data which is introduced in this paper. Because of this reason it is needed to extend the data examination techniques (for example characterization, discrimination, association analysis, classification, clustering, outlier analysis, evaluation analysis) to multi-valued data so that we get more exact and consolidated multi-valued data sets. We say that multi-valued data analysis is an expansion of the standard data analysis techniques. The objects of multi-valued data sets are represented by multi-valued attributes and they contain more than one value for one entry in the data base. An example for this type of attribute is 201C;languages known201D; .this attribute may contain more than one value for the corresponding objects because one person may be known more than one language
Sentiment Analysis and Opinion Mining from Social Media : A Review
No full textUbiquitous presence of internet advent of web 2 0 has made social media tools like blogs Facebook Twitter very popular and effective People interact with each other share their ideas opinions interests and personal information These user comments are used for finding the sentiments and also add financial commercial and social values However due to the enormous amount of user generated data it is an expensive process to analyze the data manually Increase in activity of opinion mining and sentiment analysis challenges are getting added every day There is a need for automated analysis techniques to extract sentiments and opinions conveyed in the user-comments Sentiment analysis also known as opinion mining is the computational study of sentiments and opinions conveyed in natural language for the purpose of decision making Preprocessing data play a vital role in getting accurate sentiment analysis results Extracting opinion target words provide fine-grained analysis on the customer reviews The labeled data required for training a classifier is expensive and hence to over come Domain Adaptation technique is used In this technique Single classifier is designed to classify homogeneous and heterogeneous input from di_erent domain Sentiment Dictionary used to find the opinion about a word need to be consistent and a number of techniques are used to check the consistency of the dictionaries This paper focuses on the survey of the existing methods of Sentiment analysis and Opinion mining techniques from social medi
Security Enhancement in Image Steganography
Get PDFSteganography helps in communication of secured data in several carries like images, videos and audio. It undergoes many useful applications and well known for ill intentions. It was mainly proposed for the security techniques in the increase of computational power, in order to have security awareness like individuals, groups, agencies etc. The factors that are separated from cryptography and water making are data is not detectable; capacity of hidden data is unknown and robustness of medium. The steganography provides different methods existing and guidelines. The current technology of image steganography involves techniques of LSB in image domain but once the attacker acknowledges that medium is containing embedded data he will attack the medium and breaks into the secured content. In this paper we are discussing how to protect the steganography image by embedding it into another medium using mat lab. Here we work on image matrices to perform the steganography. Lightness adjustment on the matrix is done to reduce the brighter pixels in image. The lightness decreased image then embedded into another cover image by matrix difference technique (will be discussed in detail)
Neural GDFS: Neural Network Guided DFS for Progressive Cluster Performance on Large Data Set
Get PDFGlusterFS is the most advanced Distributed File System. GlusterFS uses Devis Meyer2019;s Hashing algorithm as a one way hashing function to save file across the cluster network. Based on GlusterFS, we introduce a new kind of DFS known as NeuralGDFS. NeuralGDFS will incorporate the probabilistic neural network to identify the most probable Brick in which requested file might be located. We also studied GlusterFS performance in virtual cloud environment
Design
Get PDFMultimedia data security is very important for multimedia commerce on the internet and real time data multicast. An striking solution for encrypting data with adequate message security at low cost is the use of Scrambler/Descrambler. Scramblers are necessary components of physical layer system standards besides interleaved coding and modulation. Scramblers are well used in modern VLSI design especially those are used in data communication system either to secure data or re-code periodic sequence of binary bits stream. However, it is necessary to have a descrambler block on the receiving side while using scrambling data in the transmitting end to have the actual input sequence on the receiving end. Scrambling and De-scrambling is an algorithm that converts an input string into a seemingly random string of the same length to avoid simultaneous bits in the long format of data. Scramblers have accomplish of uses in today's data communication protocols. On the other hand, those methods that are theoretical proposed are not feasible in the modern digital design due to many reasons such as slower data rate, increasing information, circuit hazards, uncountable hold-up etc. Therefore it is requisite for the modern digital design to have modified architecture to meet the required goal. We will recommend here modified scrambler design which is perfectly suitable for any industrial design
Web Usage Mining:A Novel Approach for Web User Session Construction
Get PDFThe growth of World Wide Web is incredible as it can be seen in present days. Web usage mining plays an important role in the personalization of Web services, adaptation of Web sites, and the improvement of Web server performance. It applies data mining techniques to discover Web access patterns from Web log data. In order to discover access patterns, Web log data should be reconstructed into sessions. This paper provides a novel approach for session identification
Implementation of AES with Time Complexity Measurement for Various Input
Get PDFNetwork Security has a major role in the development of data communication system, where more randomization in the secret keys increases the security as well as the complexity of the cryptography algorithms. In the recent years network security has become an important issue. Cryptography has come up as a solution which plays a vital role in the information security system against various attacks. This security mechanism uses the AES algorithm to scramble data into unreadable text which can only be decrypted with the associated key. The AES algorithm is limited only for text as an input. It also has, the more time complexity. So it suffers from vulnerabilities associated with another type of input and time constraints. So its challenge to implement the AES algorithm for various types of input and require less decryption time. The propose work demonstrate implementation of a 128-bit Advanced Encryption Standard (AES), which consists of both symmetric key encryption and decryption algorithms for input as a text, image and audio. It also gives less time complexity as compared to existing one. At the last stage comparing the time complexity for encryption and decryption process for all three types of input. This paper also demonstrates a side channel attack on the standard software implementation of the AES cryptographic algorithm.d
Effective Detection and Prevention of Ddos Based on Big Data-Mapreduce
Get PDFDistributed Denial of Service (DDoS) attacks is large-scale cooperative attacks launched from a large number of compromised hosts called Zombies are a major threat to Internet services. As the serious damage caused by DDoS attacks increases, the rapid detection and the proper response mechanisms are urgent. However, existing security methodologies do not provide effective defense against these attacks, or the defense capability of some mechanisms is only limited to specific DDoS attacks. Therefore, keeping this problem in view author presents various significant areas where data mining techniques seem to be a strong candidate for detecting and preventing DDoS attack. The new proposed methodology can perform detecting and preventing DDoS attack using MapReduce concepts in Big Data.Thus the methodology can implement for both detecting and preventing methodologies
Enhancing the Security and Quality Image Steganography using Hiding Algorithm based on Minimizing the Distortion
Get PDFIn this paper, highest state-of-the-art binary image Steganographic approach considers the spinning misinterpretation according to the personal visual structure, which will be not secure when they are attacked by Steganalyzers. In this paper, a binary image Steganographic scheme that aims to reduce the hiding misinterpretation on the balance is presented. We excerpt the complement, turn, and following-invariant local balance arrangement from the binary image first. The weighted sum of Complement, Turn, And Following-Invariant Local Balance changes when spinning one pixel is then employed to allot the spinning misinterpretation corresponding to that pixel. By examining on both simple binary images and the composed image constructed message set, we show that the advanced appraisal can well describe the misinterpretations on both visual aspect and statistics. Based on the proposed measurement, a practical Steganographic scheme is develope
Multi Spectral Band Selective Coding for Medical Image Compression
Get PDFMedical image compression has recently evolved as an area of research for progressive transmission The distance based medical diagnosis demands for high quality imaging at faster data transfer rate As the information s are highly informative each pixel information defines a sample observation Hence the coding in medical diagnosis need to be of higher accuracy than conventional image coding In the approach of image coding multi spectral coding is developed as new coding approach to achieve the objective of higher visualization accuracy With this observation in this paper a multi spectral coding using multi wavelet transformation is developed The multi spectral coding is improved by a band selective approach using inter band correlation factor The evaluation factors for such a coding technique are observed to be improved over conventional multi-spectral codin