Volume 5 Number 2 September 2015


An Efficient Supervising Technique in Viral Hepatitis Surveillance System
S R Swarnalatha, G M KadharNawaz

Abstract: Dimension reduction is a critical data preprocessing step for many database and data mining applications, such as efficient classification of relevant and irrelevant groups in high-dimensional data. In the literature, a well-known dimension diminish algorithm is Linear Discriminant Analysis (LDA), Singular Value Decomposition (SVD). Due to the design complexity and poor efficiency, we need a design that should be able to help them to make a good decision. In this paper, we propose an LDA-based Fuzzy classifier, called Diminish Fuzzy (DF) classifier, which applied Medical Research. This method is used to reduce the risk of error in medicine field, especially Hepatitis Diagnosis. The proposed system is an intelligent system for the diagnosis of Hepatitis B virus disease. Hepatitis is one of the serious diseases which demands expensive treatment and major side effects can appear very often. The intelligent system consists of DF classifier which gives the output whether the patient is Hepatitis B positive or not and the severity of the patient. Finally we evaluate the effectiveness of the DF classifier in terms of classification error rate on the reduced dimensional space. Our experiments based on real-world data sets reveal that the classification rate achieved by the DF Classifier algorithm is better than other LDA-based algorithms.


Analysis of Classification Techniques for Mining Reviews Using Lexicon and WordNet Using R
A Sharmista, M Ramaswami

Abstract: With the exponential growth of social media i.e. blogs and social networks, organizations and individual persons are increasingly using the number of reviews of these media for decision making about a product or service. Opinion mining detects whether the emotions of an opinion expressed by a user on Web platforms in natural language, is positive or negative. This paper presents extensive experiments to study the effectiveness of the classification of English type opinions in three categories: positive, negative and none. For this study, technological products corpora have been used. Furthermore, we have conducted a comparative assessment of the analysis of two classification techniques: J48 and C50 using the effect of both Opinion Lexicon and WordNet. Experimental results shows that the WordNet based sentiment classification perform well over Opinion Lexicon based classification. The proposed technique can also be used with any other language. The whole work is implemented using ‘R’ language.


An Approach to generate Cluster based Napped Associate Template Mining using Association Rules
Dr. R. U. Anitha, M.Menakapriya, S.Nithyananth

Abstract: In this research we process a new technique called Napped Associate Template Mining (NATM) for concern forced Data Mining. It was used to find all the rules that capture the minimum support and minimum confidence constraints. In this proposed work, new template match technique to cluster association rules, based on the similar attributes, template matching clustering algorithm is used to cluster the rules. This work is used to combine more number of rules with a contingent value. Based on the contingent value, the result will be declared whether the rules or cluster or not.


White Blood Cell Analysis Using Watershed and Circular Hough Transform Technique
S Pavithra, J. Bagyamani

Abstract: In medical diagnosis, blood cell counts play very important role. The major issue in clinical laboratory is to produce a precise result for every test especially in the area of blood cell count. The number of blood cell is very important to detect as well as to follow the treatment of many diseases like anemia, leukemia etc. Blood cell count gives the vital information that helps in the diagnose of many of the patient’s sickness. The old conventional method of Blood cell counting under microscope gives an unreliable and inaccurate result depending on clinical laboratory and technician skill. This paper presents complete and fully automatic method for White Blood Cell (WBC) analysis from microscopic blood cell images. The whole work has been developed using MATLAB environment.


Cloud Security & DES Algorithm A Review
Dhina Suresh, Dr. M. Lilly Florence

Abstract: Cloud is an evolving trend today. It is an internet based service delivery model which provides internet based services, computing and storage for users. There are also disadvantages with cloud computing. Data security and privacy protection issues remain the primary problem in cloud. It is required to protect the stored data and applications in the cloud. This article discusses on the basics of cloud and the security issues in it. It gives a note on the existing cryptographic algorithms and it gives a detailed discussion on the DES algorithm.


Improving the Navigability of Web Pages using Genetic Algorithm
N. Thangammal, Dr. A. Pethalakshmi

Abstract: With the recent spurt in the number of engineering web applications, the challenge of improving the quality of the engineered applications has come to the forefront. One important aspect of the quality of a web application is navigability. It represents the ease and convenience with which the user of the web site can have quick access to all the information contained in the web site. Genetic algorithms that mimic the natural process of evolution to uncover solutions to problems have achieved huge success with complex problems across multiple domains. This paper presents a genetic algorithmic approach to improving the navigability of web pages that identifies the best sequence of transformations that can improve navigability.


FP-AODV Forwarding in Mobile Adhoc Network
K. Veeramani, Dr. I. Laurence Aroquiaraj

Abstract: In this paper, we focus upon the increase the throughput of an on-demand distance vector routing (AODV) protocol for mobile and wireless ad hoc networks. We propose to FP-AODV protocol increases the packet delivery ratio, throughput better than other protocols. And decrease the end to end delay and packet dropping. AODV protocol is extended with a plunge factor that induces a randomness feature to result in Finest Path Selection Ad-Hoc On-Demand Routing (FP-AODV) protocol.


Optimal Data Prediction and Classification Applicable for Intelligent Heart Disease Diagnosis System
K. Jayavani, G. M. KadharNawaz

Abstract: In the past years, medical data mining has become a popular data mining subject. Researchers have proposed several tools and several methodologies for developing effective medical expert systems. Diagnosing heart diseases is one of the much need topics and many researchers have tried to develop intelligent medical expert systems to help the physicians. In this paper, we proposed a novel, particle separable optimization Algorithm, which is derived from particle swarm optimization and efficient nearest algorithm. Efficient Nearest neighbor (ENN) is very simple, most popular, highly efficient and effective algorithm for pattern recognition. ENN is a straight forward classifier, where samples are classified based on the attributes class of their nearest neighbor. Medical data bases are high volume in nature. If the data set contains redundant and irrelevant attributes, classification accuracy will be degraded. The main objective of this approach is to extract rules for diagnosis the existence or in existence of heart disease in a patient. The PSO composed of three steps, first to extract the feature from the database using pso, second one is classify the feature based on efficient nearest neighbor search, finally the optimization level measured from polynomial and Fourier series regression method. In addition classification accuracy improved by this method compared with conventional PSO.


Exploring Highly Structure Similar Protein Sequence Motifs using SVD with Soft Granular Computing Models
E Elayaraja, K Thangavel

Abstract: Vital areas in Bioinformatics research is one of the Protein sequence analysis. Protein sequence motifs are determining the structure, function, and activities of the particular protein. The main objective of this paper is to obtain protein sequence motifs which are universally conserved across protein family boundaries. In this research, the input dataset is extremely large. Hence, an efficient technique is demanded. A Rough Granular computing model is created to efficiently extracting protein motif data that transcends protein families. Before apply this model, the very first step of this research is trying to reduce segments. The literature suggests that the Singular Value Decomposition (SVD) computing technique is more suited for reducing segments. After that the reduced segments are followed by applying Rough Granular computing model. The effectiveness of final results effectiveness is tested by several measures. The experimental results suggest that the SVD with Rough Granular computing model generates more number of highly structured motif patterns.


Performance Analysis For Blood Vessel Segmentation In Hypertensive Retinopathy
B. Saranya, Dr. I. Laurence Aroquiaraj

Abstract: Hypertensive Retinopathy (HR) is one of the most common retinal diseases. Normally, the high blood pressure will affect the various organs of the biological system, one of them is eye. The retinal blood circulation is affected, due to the damages caused by the high blood pressure; this disease is termed as the Hypertensive Retinopathy. This disease is diagnosed from the retinal images, specifically from the blood vessels of the retina. Some of these measures for finding HR using digital image are Segmentation of blood vessels, measurement of tortuosity, diameter measurement, finding the artery vein ratio. A method is proposed for segmenting the blood vessels which is performed in various steps. The footprint of the blood vessels is more visible in the gray scale image and green extracted image. The noise in the image is removed using Gabor Wavelet and Sharpening Spatial filters. The segmentation is performed based on the approximation so the Rough Entropy is well suits to this retinal image to distinguish the blood vessels from the other parts. The Evaluation measure such as Probability Rand Index (PRI), Global Consistency Error (GCE), Local Consistency Error (LCE) measures are used for comparison process between gray scale image and green extracted image which gives better segmentation result.