Volume 4 Number 2 September 2014
1 |
Imporovement in Utilization of the Spectrum using Cognitive Radio nodes Abstract: Currently, the research towards spectrum management has been increased inorder to avoid the scarcity of the spectrum and to improve the utilization of the spectrum which shows that researchers concentration towards spectrum utilization also get increased. This paper provides how ways of spectrum utilization and implementation varies from researchers to researchers. The main objective of the paper is to improve the overall spectral efficiency by sharing the spectrum among the service providers and avoiding the spectrum scarcity. The proposed system validates the utilization of spectrum sharing in three different ways like Normal utilization, applying Fuzzy logic system and predicting traffic pattern and finally proves that utilization of the spectrum has been improved with reduced call blockage, reduced interference and reduced high traffic patterns of the calls. This paper also gives the study of Spectrum sharing with different technologies. |
2 |
Naive Bayes Classification Technique for Analysis of Ecoli Imbalance Dataset Abstract: The classification technique is a systematic approach to build classification models from an input data set. The techniques include rule-based classifiers, decision tree classifiers, support vector machines, neural networks and Naive Bayes classifiers. Every technique employs a learning algorithm to discover a model that best fits the relationship among the attribute set and class label of the input data. The model generated by a learning algorithm should both fit the input data well and correctly forecast the class labels of records it has never seen before. Therefore, a key objective of the learning algorithm is to construct models with good generality capability. That is the models that accurately predict the class labels of previously unknown records. In this paper we are analyzing the performance of 3 classifiers algorithms namely Naïve Bayes, Instance Based K-Nearest Neighbor (IBK) and J48 Decision Tree. From the experimental results, it is found that Naïve Bayes technique performs better than the other two techniques. We use the ecoli protein datasets for calculating the performance by using the cross validation parameter. And finally we find out the comparative analysis based on the performance factors such as the classification accuracy and execution time is performed on all the algorithms. |
3 |
An Identification of Performance Ccuracy Gain Under Realistic Scenario Abstract: Cooperative Positioning (CP) in VANET is mainly for road safety applications such as cooperative collision warning system etc. CP is an approach for location determination within wireless adhoc sensor networks. The goal of CP is to allow neighbor nodes to work together to collectively improve the accuracy of their positions. Although this technique is well known, the efficiency of CP under real world scenario is not considerable. So in our paper, we propose a technique to increase the efficiency of CP. This technique includes the formation of range vector, extended range matrix and also calculates the accurate position of a vehicle using the trilateration method. Our results demonstrate that, even under dense traffic conditions, these protocol improvements achieve a twofold reduction in packet loss rates and increase the positioning accuracy of CP by 10-15%. |
4 |
An Experimental Analysis of Evolutionary and Swarm Intelligence Algorithms for 3D HP Structure Prediction Abstract: Predicting the structure of protein has been the focus of the scientific research, but it has challenging in bioinformatics due to the computational complexity. The protein structure is determined by the experimental methods such as X-ray crystallography and nuclear magnetic resonance (NMR) spectroscopy. These methods cannot always be applied. So, the computational methods are frequently used to predict the structure with lowest free energy conformations. The lowest free energy is calculated based on the hydro polar and hydrophilic interactions. Many of the computational algorithms are used to solve the protein structure problem. In this comparative study the evolutionary algorithm, Genetic Algorithm and swarm intelligence algorithms Ant colony optimization (ACO) and artificial bee colony (ABC) algorithms are used and comparison is based on its energy value. The lowest energy value can easily predict the well known structures. |
5 |
A Novel Feature Selection Algorithm for Heart Disease Classification Abstract: Humans are affected by different life threatening diseases. One among them is heart disease. Medical practitioners pay more attention to this disease. Patients undergo different diagnostic procedures to identify the factors related to the disease. The results of the procedures yield different variations of the disease. However, there may be relevant, redundant and irrelevant features representing the disease. Identifying the relevant features is considered as an optimization problem. Computational intelligence techniques have been widely applied to determine the relevant features for disease classification. This paper uses a metaheuristic algorithm to determine the optimal feature subset with improved classification accuracy in heart disease diagnosis. A Binary Artificial Bee Colony (BABC) algorithm is used to find the best features in the disease identification. The fitness of BABC is evaluated using K–Nearest Neighbor (KNN) method. Results are validated using Cleveland Heart disease dataset taken from the UCI machine learning repository. The results indicate that, BABC–KNN outperform the other methods. |
6 |
Binary Decision Tree Classification based on C4.5 and KNN Algorithm for Banking Application Abstract: In current era, database is widely used for storage purpose. History of these data which are stored in database and data warehouse has to be used optimally to analysis and predict the current trends. Mining of the data is required to perform the analysis. Data mining extracts the knowledge from the database or data warehouse. Extracted knowledge is represented in the form of various models. Among the models in data mining, Classification is the widely used model of data representation. To classify the data in the dataset, decision tree approach is introduced. There are various algorithms introduced in data mining technologies of which C4.5 identified to be a famous algorithm. C4.5 classifier uses the information gain ratio as the parameter to build the decision tree. Rules are extracted from the decision tree. Inconsistencies in the dataset are also considered before decision tree is built. KNN algorithm is used to resolve the inconsistencies due to missing data. This KNN algorithm is a clustering based approach. Size of the decision tree will depends on the attribute types at each node namely categorical or numerical. Thus split at a node will lead to numerous child nodes depending upon the type of attributes at the node to be split. This paper proposes a binary decision tree construction irrespective of attribute types. Thus the project always built the binary decision tree with only two splits at each attribute. The modified decision tree also proves the efficiency in terms of true positive rate and false positive rate compared to initial decision tree. |
7 |
Anomaly Detection using Clusters and Proximities Measures Abstract: In an increasing number of security issues, intruder detection system are used to detect an insecure network attacks. There are so many attacks, in real time process; to detect it some of IDS system is used for filtering such data packets. This paper analysis the anomaly based intrusion detection techniques. AIDS is a system for detecting intrusions, type of attacks that falls out of normal process system activity and classifying it as either normal or ano malous. Anomaly detection searches for an unusual cases based on behavior analysis deviations. It quickly detects the attack in data analysis process by clustering. A StepWiseClustering (SWC) algorithm is used to detect the attack in unusual cases. |
8 |
Matrix Based Key Pre- Distribution Scheme for Wireless Sensor Networks Abstract: Wireless sensor networks are used in various applications now-a-days. As they are deployed in open area, there is a need for key management in order to protect the information stored in sensor nodes. To address this problem, we use key pre-distribution scheme. In this paper, we propose a new scheme based on symmetric matrix and maximum rank distance (MRD) codes where the size of the symmetric matrix is kept constant to reduce the memory requirement at each node. Some information about the matrix G and matrix A is stored in sensor nodes to generate secret key between them and for secure communication. Only two messages are required to generate a secret key between two nodes and thus it reduces the communication overhead. This scheme has greater network connectivity and scalability. Newly deployed nodes can generate a key without changing any information on previously deployed nodes. To provide additional security, the final result from key generation scheme is applied to the division remainder hash function and the resultant value is used as the secret key between the nodes. |
9 |
Evaluation of Wavelets and Classifiers in Classifying Cardiovascular Disorders using Wavelet Transform Abstract: The cardiac abnormalities are to be detected and treated earlier in order to avoid its adverse effects. The PQRST properties from the recorded ECG are used to analyze the type of arrhythmia. Noise removal has to be done effectively for further processing of the ECG signals. In this paper, twelve different ECG samples from MIT BIH Arrhythmia database are analyzed using six mother wavelet functions- haar, db8, sym5, coif5, bior4.4 and rbio4.4. And cardiac disorders like Myocardial Infarction, Premature Ventricular Contraction (PVC), Ventricular Tachycardia, Supra ventricular arrhythmias, ST deviation, Ventricular Fibrillation (VF) are classified using Naive Bayes (NB) classifier and Support Vector Machine (SVM) classifier. The databases are extracted from MIT-BIH, EURO, VFDB, SVDB, MIT-ST Change, CUVTDB databases. Wavelet transform algorithm is used for extracting the features. The wavelets are evaluated using three different performance measures such as Peak Signal to Noise Ratio, Mean Squared Error and Mean Absolute Error. The experimental results shows that coif5 wavelet is efficient in significantly reducing the baseline wandering in the ECG signals and the samples are classified using Naive Bayes classifier and Support Vector Machine classifier which achieved 94.45% and 96.76% accuracy respectively. |
10 |
An Analysis on the Main Factors of Occupational Stress among Indian Women – A Soft Computing Approach Abstract: Work and family are the two most important aspects in women’s lives. Balancing work and family roles have become a key personal and family issue for many societies. There are many facets in working mother’s lives that subject to stresses. They deal with home and family issues as well as job stress on a daily basis. Imbalance between work and family life arises due to a number of factors. Various factors appear to strengthen the brunt of pressure on women. Stress experienced by women at a workplace affects not only their professional life, but also family life and social intercourses. In the women’s opinion, an unpleasant workplace is such a workplace where the feeling of mental workload is connected with the lack of rewards (motivation), uncertainty resulting from organization of daily chores and lack of support from others. The high general level of stress was noted among the group of women working in IT sectors as well as among those who perform physical work (seamstresses). No significant dependencies were concluded between socio-demographic variables and the general level of exposure to intensified stress in the examined professional groups. The above research confirms the need for further examination of the working environment of women and its impact on health. Obviously, attempts should be made in order to improve the conditions of work for women, bearing in mind the fact that the adoption of neutral attitude towards the sexes when assessing risk and undertaking preventive activities may result in the female gender being underestimated or even disregarded. Regarding to woman’s household tasks and families responsibility, it is important to measure the level of occupational stress in working women and assess relation between occupational stress and family difficulties in order to obtain knowledge for health care providers to provide support to the working women and their families. There is a great need for research on working women, especially concerning the impact and occu rrence of job stress on children mental health. Longitudinal data on young childbearing families are needed to examine the complex issues of work-family surrounding the family in today’s world. The purpose of this analysis is the evaluation of women’s exposure to stress-inducing factors at work and definition of a scale of the problem. The research was conducted on different professional groups of women. The research tool was the standardized questionnaire for Subjective Work Evaluation. The raw result was obtained on the basis of summing up all the points. The results of stress factors were quoted as mean results of raw values and were referred to results defined as high for a given factor. The results obtained were analyzed using a soft computing technique. |