International Journal of Intelligent Engineering and Systems

1

Authors: Mohcine Bouksim, Fatima Rafii Zakani, Khadija Arhid, Mohamed Aboulfatah, Taoufiq Gadi

In the last decades, 3D models have known a significant expansion, as we can see them in many fields such as computer vision, biology, augmented reality, engineering, and even medicine. They become more and more popular and accessible than ever. Due to these facts, there is an increasing need for a retrieval system, which can return the most similar 3D objects existing in the database to the query. In the present paper, a new retrieval approach is proposed, based on a multi-criteria method to generate a compact descriptor, which represents a signature for each 3D model. The key idea behind this approach is to try to extract the best out of each criterion (i.e., measure) by extracting a combined score using the Data envelopment analysis method (DEA). The results of the proposed method are very satisfactory and outperform some commonly used retrieval methods which highlight the potential and the performance of our approach.

20

Authors: Guchenko Mykola, Sokhin Natalia, Lugovyj Oleksandr, Dobiáš Václav, Čejka Jiří, Ludvík Miroslav

The actual task is to research QoS (Quality of Service) instruments in order to improve existing and develop new methods. This paper presents the modified method of double QoS, which differs from standard one by application of multi-level traffic analysis by user groups, types of traffic and priorities. The efficiency of the proposed method is proved experimentally on routers with the Mikrotik operating system by means of local computer network simulation, based on virtual machines. The experiments were conducted with generated multi-traffic with a uniform distribution for Packet Size and Internet Departure Time parameters. Compared results of router work with standard and modified methods. Using the proposed method allows reducing the number of dropped packets by simply configuring the router system tools and thus improving the quality of user service.

41

Authors: Syed Muzamil Basha, Dharmendra Singh Rajput, Vishnu Vandhan

Boosting is the method used to improve the accuracy of any learning algorithm, which often suffers from over fitting problem, because of inappropriate coefficient associated to the data points. The objective of our research is to train the data, such that the weighing error of linear classifier goes to zero and classify the sentiments accurately. In this paper, Gradient ascent approach is used to minimize the weighing error of sentiment classifier by predicting the proper coefficients to the data points in trained dataset. When compared to previous studies on designing a strong classifier, our research is novel in the following areas: Estimation of Maximum Likelihood for logistic regression using Gradient ascent and making use of weights of metric in understanding the behavior of AdaBoost algorithm in classifying the sentiments. In our finding, the first decision stump has training error of 30.44%. After thousand iterations, we observed a smooth transition, where the classification error tends to go down to 8.22% and actually stays at same value. Finally, concluding that Boosting algorithm outperforms Random Forests with lesser Mean squared Test Errors.

62

Authors: Syaifulnizam Abd Manaf, Norwati Mustapha, Md Nasir Sulaiman, Nor Azura Husin, Helmi Zulhaidi Mohd Shafri, Mohd Norhisham Razali

Observation satellites orbiting the Earth provide important surveillance information that helps in identifying various types of land cover. As such, the resolution of satellite images is critical to ensure high accuracy in classifying land cover types. Clearly, high-resolution images are desirable, but such images are prohibitively expensive. Hence, the use of medium-resolution satellite images seems more economical and practical. Several techniques have been developed to monitor the conditions of land covers across the world, such as aerial photography, ground survey, and remote sensing. Among the three techniques, remote sensing is the best, given its ability to monitor vast geographical areas more accurately and much faster compared to those of the other two techniques. In recent decades, many countries have been relying on remote sensing to monitor the conditions of coastal areas by extracting shorelines from satellite images. To date, several pixel-based methods have been proposed for the extraction of shorelines, but most of these methods are fraught with problems. Thus, the object-based approach is proposed using a combination of segmentation algorithms, namely Felzenswalb, Quickshift, and SLIC, together with 15 machine learning classifiers, to classify segmented images of Langkawi Island. The performance of the segmentation algorithms and machine learning classifiers were assessed in terms of segmentation time and overall accuracy in four experimental settings comprising of three different parameters. The research findings showed that the proposed hybridization of SLIC segmentation and Extra Tree classifier was the most efficient and accurate technique compared to other combinations of techniques in extracting the shoreline of the study area. Specifically, SLIC was faster than Felzenswalb and Quickshift by as much as 37 times and 500 times, respectively. Together with Extra Tree classifier, SLIC managed to achieve 100% overall accuracy in the object-based classification in three out of four settings compared to the rest of the techniques tested in the study.

83

Authors: Clara Kanmani Arulanandu, SumaVasu Deva Murthy, Guruprasad Nagraj

In this paper, a new secured data model for cloud computing is proposed which uses partial resource description framework (RDF) encryption and token based access control system in which sensitive data in an RDF graph is encrypted and all other non-sensitive data are publicly lucid. The security process, the decryption process, and query process are the three essential procedures in this framework. The consequence of the security process is the encrypted data, encrypted metadata, and plain text fragments. The proposed technique permits the token based access control system for the decryption procedure. The query process incorporates the map reduce framework is for lessening the immense measure of employments. At long last, the query answer is sent to the user in light of the access token list (AT-list) of the system administrator. Our test comes about demonstrate that, the performance of the proposed technique is assessed in view of the precision, recall and execution time of the framework. Our proposed approach is actualized using Java and keep running on Windows XP framework and the Lehigh University Benchmark (LUBM) datasets are used for our examination. In the paper this new secured RDF data model is deployed and tested using AWS elastic beanstalk.

104

Authors: Dhanavijayan Ramakrishnan, Sudhakar Pushparajan

This paper demonstrates an ideal procedure to find the desirable bus location and operating power of distributed generation in radial power distribution system with the aid of unified particle swarm optimization algorithm. In general, an electrical power network economics primarily relies on the conductor line losses. Hence in the proposed study, the network active power loss is picked as the minimization problem, and it is reduced by incorporating optimum distributed energy source at appropriate buses using proposed unified particle swarm optimization algorithm. The system power loss is handled as the cost function for each particle in a swarm. The simulation analysis of the purposed procedure is successfully executed for different radial distribution system using MATLAB. The prosperous solutions achieved from the simulation studies manifest that the high level of system loss reduction and desirable bus voltage profile, when compared against earlier existing methods such as genetic algorithm, particle swarm optimization, and parameter improved particle swarm optimization.

121

Authors: Swati Srivastava, Bipin Kumar Tripathi

This paper aims to develop a recognition system which incorporates a novel fusion from different areas of computational intelligence along with a proposed artificial neuron structure. Our Deep Hybrid Computational Model (DHCM) is a novel aggregation of fuzzy clustering fused with evolutionary searching and a neural network based on a proposed neuron model TROIKA. The deep architecture with evolutionary fuzzy clustering followed by hybrid neural network, built upon novel neuron TROIKA, yields a powerful tool for face recognition applications. The proposed TROIKA neuron is based on nonlinear aggregation functions which enables our hybrid neural classifier to inculcate the benefits of its computational power resulting in superior learning and generalization. The proposed TROIKA neuron reduces the complexity of DHCM because very few neurons are sufficient to recognize a subject in database. The experimental results completed on two benchmark face datasets-INDIAN and FERET datasets demonstrated the effectiveness of the proposed model. The performance comparisons revealed the outperformance of (i) TROIKA based model over conventional neuron based model (ii) proposed model over other classification methods.

142

Authors: Jayapal Joseph, Ravi Subban

Nowadays, the exploitation of images is inevitable in all the domains. However, the major issue being faced by the images is ‘noise’, which seriously affects the image quality. In order to address this issue, this article presents a novel denoising algorithm that is based on superpixel clustering and dictionary learning process. The superpixel clustering groups the similar pixels together to form a superpixel. As soon as the superpixels are computed, the sparse coefficients are computed by the steepest descent orthogonal matching pursuit algorithm. The process of dictionary updation is achieved by Discriminative K- Singular Value Decomposition (DK-SVD). The dictionary is trained by DK-SVD algorithm, as the discriminative capability of the algorithm is greater. The proposed approach shows promising results, which when compared to the existing techniques. Additionally, the time consumption of the proposed approach is very minimal than the comparative techniques.

163

Authors: Metta Venkata Satya Sairam, Mupparaju Sivaparvathi

Radio spectrum (3 KHz to 300 GHz) is a limited natural resource and efficient utilization of the spectrum can be accomplished by cognitive radio (CR). In CR, spectrum sensing is one of the significant challenging issues. Cooperative spectrum sensing (CSS) based CR is a captivating solution to enhance spectrum sensing. In CSS, multiple secondary users (SUs) report their results to fusion center (FC) and final decision is taken on the status of the primary user (PU) based on the predetermined fusion rule. This cooperative decision in CSS based CR, improves the spectrum sensing performance. Nevertheless, the foremost drawback in CSS is low throughput due to reporting time introduced by the multiple SUs. We propose a novel method to reduce the reporting time so as to enhance the throughput performance for the same number of SUs in CSS based CR. In the proposed scheme k SUs, where k is even, are divided into two groups, each with k/2 users. Two groups report their sensing results to two FCs separately and final decision is taken by the third FC. Consequently, the reporting time is condensed in the proposed scheme. The sensing performance of the proposed CSS (PCSS) for all possible fusion rules is reviewed. Analytical expressions for probability of false alarm, probability of detection and throughput are derived for PCSS. Both simulation results and theoretical analyses validate that PCSS improves throughput significantly when compared with CSS.

181

Authors: Bukka Narendra Kumar, Mantena S. V. Sivarama Bhadri Raju, Bulusu Vishnu Vardhan

With the huge development of the usage of computer over network and advancement in applications running on various platforms captures the attention towards network security. The Intrusion Detection System (IDS) plays a vital role in detecting anomalies and attacks in the network. Earlier approaches of IDS relied on Machine Learning (M L) techniques. Due to some limitations, a better approach is needed. A combination of Machine Learning techniques and data preprocessing is an effective approach for IDS. In this work, a new dimensionality reduction technique combined with the Multi-class SVM (Support Vector Machine) is proposed for intrusion detection. In the proposed model, Multi-Linear Dimensionality Reduction (ML-DR) is proposed as a feature extraction technique to reduce the dimension in order to shorten the training time. A Multi-class SVM (M-SVM) is used to detect whether the action is an attack or not. Here the Multi-class SVM is adopted to perform multi-attack classification in a layered fashion. Radial Basis Function Kernel is used as a SVM kernel. NSL-KDD data set is used for the performance evaluation of the proposed approach. The performance metrics such as classification accuracy, false alarm rate and the correlation coefficient are evaluated to measure its efficiency. In comparison with other detection approaches, the experimental results show that the proposed model outperforms the higher classification accuracy.

201

Authors: Vulapula Sridhar Reddy, Barige Thirumala Rao

In this paper, we plan a combined clustering and geometric data perturbation approach for improving privacy preservation of health care data in hybrid clouds. We will possibly plan an answer that can productively give protection to information put away in the cloud without presenting substantial overhead on both computation and communication. At first, the high-dimensional data are separated into various parts by utilizing the K-mean clustering method, each partition is considered as a cluster. At that point the mean estimate of each cluster processed; after that the contrast between the each cluster member and the mean of the cluster value is computed. In the next stage, the clustered information is perturbed by utilizing the Geometric Data Perturbation (GDP) algorithm which makes the values difficult to be recognized. These perturbed values are stored in the public cloud and the key parameters for randomizing and clustering is stored in the private cloud. Our approach would contribute in reduction of too much storage on private cloud on the off chance that we basically store the entire sensitive information on private clouds. The experimental results show that, the GDP algorithm has better privacy preserving compared with the other existing methods.

221

Authors: Megala Tulasiraman, Vivekanandan Kalimuthu

Regression testing is one among the strongest testing criteria which ensure the quality of the software under test. However, regression testing is too expensive due to execution of too many test cases .Test case prioritization is one of the traditional techniques which improve the regression testing by proposing a test case order that increases the rate of fault detection . In this paper, we propose a cost – cognizant history based test case prioritization approach that utilizes the historical information of the test cases like cost, fault identified by the test case and the severity of the identified fault for prioritization. Also an artificial immune system based, Clonal selection algorithm is proposed to find an effective test case order from existing test suite. To evaluate the proposed approach, controlled experiment were performed and the evaluation results indicates the test case order produced by proposed approach shows improvement in terms of average percentage of fault detected per cost.

239

Authors: Biswa Mohan Acharya, Ajit Kumar Nayak

Wireless multimedia sensor network (WMSN) deals with transmitting multimedia traffic over a long range in multi-hop along multiple paths. This is so for achieving the objectives of Quality of Service (QoS). In clustered routing, the load on the cluster head nodes increase and the energy consumed at the cluster head nodes lead to energy depletion of that node which degrades the network performance. The proposed protocol focuses on QoS and energy efficiency in WMSN by considering the hierarchical deployment of heterogeneous nodes. This protocol makes use of hierarchical structure of powerful cluster head nodes and the optimized multiple paths to support reliable and energy efficient multimedia transmission in WMSN. The reliability and life of the network increases because of the fault tolerance and load balanced feature of the protocol. Simulation results describe a significant performance improvement in terms of energy consumption of the nodes and the QoS parameter.

259

Authors: Calduwel Newton Pitchai, Nismon Rio Robert

Routing in Mobile Ad-hoc Networks (MANETs) is a very challenging task since, nodes are mobile and links are frequently created and broken. Traditional proactive routing algorithms initiate path discovery process from beginning after link failure occurrence. Generally, link failure occurs due to frequent topology change, fading, multiuser interference, etc., In MANETs link stability is the important criterion for link failure. Once the link meets with failure, it affects the overall performance of MANETs topology and it is very difficult to select optimum path as it leads to more transmission delay. The objective of this paper is to develop a mathematical model for Path Selection Algorithm based on Hamming Distance and delay parameters (PSA-HD). In this paper, link failure is identified, stable link is selected from the available links to reach the destination and it determines a stable path. The simulation result reveals that there is a decrease in average End-to-End delay and increase in Packet Delivery Ratio (PDR).

276

Authors: Tamilarasan Santhamurthy, Kumar Parasuraman

The cognitive radio (CR) offers keen solution to the serious scarcity of spectrum in wireless communication. In Multi-Input and Multi-Output (MIMO), primary user recruits some secondary users to cooperatively relay the primary traffic. It is very difficult to assign channel allocation and resource distribution in heterogeneous networks. We propose EPBDRA-CRN technique which provides priority scheduling scheme resulting in more reliability and efficient data transmission. The resource allocation can be done dynamically after ensuring that the nodes satisfy the specified constraints such as fairness, minimum-rate-guarantee, and minimum-delay-guarantee. Proposed EPBDRA-CRN algorithm does resource allocation and scheduling for PU and SU to exploit the performance of entire network by reducing the interference at PU. The proposal also supports different rates of transmission with priority, minimizes the transmission delay and improves fairness and provides significant network throughput and avoids collision. We compare EPBDRA-CRN with existing DRA-CRN algorithm using ns-2 simulation. The simulation result shows that the proposal is 27% more efficient than the existing DRA-CRN algorithm.