May 2014, No.3
Print E-mail

A Software Tool for Automatic Generation of Neural Hardware

Leonardo Reis, Luis Aguiar, Darío Baptista, and Fernando Morgado-Dias
Madeira Interactive Technologies Institute and Competence Centre for Exact Sciences and Engineering, University of Madeira, Portugal

 
Abstract: Natural neural networks greatly benefit from their parallel structure that makes them fault tolerant and fast in processing the inputs. Their artificial counterpart, artificial neural networks, proved difficult to implement in hardware where they could have a similar structure. Although, many circuits have been developed, they usually present problems regarding accuracy, are application specific, difficult to produce and difficult to adapt to new applications. It is expected that developing a software tool that allows automatic generation of neural hardware while using high accuracy solves this problem and make artificial neural networks a step closer to the natural version. This paper presents a tool to respond to this need: A software tool for automatic generation of neural hardware. The software gives the user freedom to specify the number of bits used in each part of the neural network and programs the selected FPGA with the network. The paper also presents tests to evaluate the accuracy of the implementation of an automatically built neural network against Matlab.


Keywords: Artificial neural networks, feedforward neural networks, system generator, matlab, xilinx, simulink, integrated software environment.
 
  Received January 25, 2012; accepted March 28, 2013
 

Full Text

 
Print E-mail

Integrating Global and Local Application of Naive Bayes Classifier

Sotiris Kotsiantis
Department of Mathematics, University of Patras, Greece
 
 
Abstract: Naive Bayes algorithm captures the assumption that every attribute is independent from the rest of the attributes, given the state of the class attribute. In this study, we attempted to increase the prediction accuracy of the simple Bayes model by integrating global and local application of Naive Bayes classifier. We performed a large-scale comparison with other attempts that have tried to improve the accuracy of the Naive Bayes algorithm as well as other state-of-the-art algorithms on 28 standard benchmark datasets and the proposed method gave better accuracy in most cases.


Keywords: Naive Bayes Classifier, data mining, machine learning.
 
Received July 30, 2011; accepted February 28, 2013

  

Full Text

 
Print E-mail

Time Stamp Based ECC Encryption and Decryption

Addepalli VNKrishna
Department of Computer Science and Engineering, Pujya Shri Madhavanji College of Engineering and Technology, India
 

Abstract: Elliptic Curve Cryptography provides a secure means of exchanging keys among communicating hosts using the Diffie Hellman Key Exchange algorithm. Encryption and Decryption of texts and messages have also been attempted. In the paper on Knapsack over ECC algorithm, the authors presented the implementation of ECC by first transforming the message into an affine point on the EC, and then applying the knapsack algorithm on ECC encrypted message over the finite field GF(p). The knap sack problem is not secure in the present standards and more over in the work the authors in their decryption process used elliptic curve discrete logarithm to get back the plain text. This may form a computationally infeasible problem if the values are large enough in generating the plain text. In the present work a new mathematical model is used which considers the output of ECC algorithm, a variable nonce value and a dynamic time stamp to generate the cipher text. Thus by having key lengths of even less than 160 bits, the present algorithm provides sufficient strength against crypto analysis and whose performance can be compared with standard algorithms like RSA.

Keywords: ECC, Time stamp, nonce value, mathematical model.

  Received August 23, 2011; accepted March 11, 2013
  
 
Print E-mail

Edge Detection Based on the Newton Interpolation’s Fractional Differentiation

Chaobang Gao1,2, Jiliu Zhou2,3, and Weihua Zhang3
1College of Information Science and Technology, Chengdu University, China
2Key Laboratory of Pattern Recognition and Intelligent Information Processing, Chengdu University, China
3School of Computer Science, Sichuan University, China

 
Abstract: In this paper, according to the development of the fractional differentiation and its applications in the modern signal processing, we improve the numerical calculation of fractional differentiation by Newton interpolation equation, and propose a new mask, the Newton Interpolation’s Fractional Differentiation (NIFD). Then we apply this new mask to image edge detection and can obtain the better edge information image. In order to get continuous and thin edges, we synthesize a new gradient and adopt the non_maxima suppression method. For a comparison, we consider the edge map yielded by the Sobel operator and Canny operator. By contrast, we discover that the edge image obtained by NIFD operator is better than those of Sobel and Canny operators, and specially for a noisy image, NIFD operator has the best anti-noise ability.


Keywords: NIFD operator, edge detection, newton interpolation, fractional differentiation.
 
Received September 4, 2011; accepted May 22, 2012
  

Full Text

 
Print E-mail

LSSVM Parameters Tuning with Enhanced Artificial Bee Colony

Zuriani Mustaffa and Yuhanis Yusof
School of Computing, University Utara Malaysia, Malaysia

 
Abstract: To date, exploring an efficient method for optimizing Least Squares Support Vector Machines (LSSVM) hyper-parameters has been an enthusiastic research area among academic researchers. LSSVM is a practical machine learning approach that has been broadly utilized in numerous fields. To guarantee its convincing performance, it is crucial to select an appropriate technique in order to obtain the optimized hyper-parameters of LSSVM algorithm. In this paper, an Enhanced Artificial Bee Colony (eABC) is used to obtain the ideal value of LSSVM’s hyper parameters, which are regularization parameter, γ and kernel parameter, σ2. Later, LSSVM is used as the prediction model. The proposed model was employed in predicting financial time series data and comparison is made against the standard Artificial Bee Colony (ABC) and Cross Validation (CV) technique. The simulation results assured the accuracy of parameter selection, thus proved the validity in improving the prediction accuracy with acceptable computational time.

Keywords: ABC, LSSVM, financial time series prediction, parameter tuning.
 
Received November 1, 2011; accepted May 22, 2012
  

Full Text

 
Print E-mail

Fast Computation of Accurate Pseudo Zernike Moments for Binary and Gray-Level Images

Khalid Hosny
Faculty of Computers and Informatics, Zagazig University, Egypt
 

Abstract: A new method is proposed for fast computation of accurate pseudo-Zernike moments for binary and gray level images. These orthogonal moments are computed as a linear combination of accurate geometric and radial geometric moments which are computed by mathematical integration of the monomial polynomials over digital image pixels. The proposed method is fast, accurate, simple and easy programmable. A comparison with the existing methods is performed. The obtained results explain the efficiency of the proposed method.


Keywords: Pseudo-Zernike moment, geometric moments, symmetry property, fast computation, binary images, gray level images.
 
Received September 20, 2011; accepted May 22, 2012
  

Full Text

 
Print E-mail

A Hybrid Algorithm to Forecast Enrolment Based on Genetic Algorithms and Fuzzy Time Series

Haneen Talal Al-wazan1, Kais Ismail Ibraheem2, and Abdul Ghafoor Jasim Salim1
 1Department of Mathematics, Mosul University, Iraq
2Department of Computers Sciences, Mosul University, Iraq
 
Abstract: In this paper, we proposed a hybrid algorithm to forecast enrolment based on fuzzy time series and genetic algorithms, the proposed algorithms presents a good forecasting result with higher accuracy rate. Historical enrolment of the University of Alabama from year 1948 to 2010 are used in this study to illustrate the forecasting process.


Keywords: Fuzzy time series forecasting, genetic algorithms prediction.
 
Received July 25, 2012; accepted January 15, 2013
  

Full Text

 
Print E-mail

Effective Image Retrieval Based on an Experimental Combination of Texture Features and Comparison of Different Histogram Quantizations in the DCT Domain

Fazal Malik and Baharum Baharudin
Computer and Information Sciences Department, Universiti Teknologi PETRONAS, Malaysia

 
Abstract: The compressed domain is appealing for the image retrieval because of the direct efficient feature extraction; moreover, currently almost all the images are available in a compressed format using the Discrete Cosine Transformation (DCT). In this paper, the quantized histogram statistical texture features are extracted from the DCT blocks using the significant energy of the DC and the first three AC coefficients of the blocks and are used for the retrieval of the similar images. The effectiveness of the image retrieval is analyzed by performing an experimental comparison of the different combinations of the texture features to get an optimum combination and the comparison of the different quantization bins by using the optimum combinations of the features. The proposed approach is tested by using the Corel image database and the experimental results show that the proposed approach has a robust image retrieval using the combinations of the features with the different histogram quantization bins in the frequency domain.

Keywords: Compressed domain, feature extraction, DCT, statistical texture features, quantized histogram.
 
Received July 10, 2012; accepted January 16, 2013
  

Full Text

 
Print E-mail

Reduct Algorithm Based Execution Times Prediction in Knowledge Discovery Cloud Computing Environment

Kun Gao, Qin Wang, and Lifeng Xi
Computer and Information Technology College, Zhejiang Wanli University, China

 
Abstract: Cloud environment is a complex system which includes the matching between computation resources and data resources. Efficient predicting tasks execution time is a key component of successful tasks scheduling and resource allocation in Cloud Computing Environment. In this paper, we propose a framework for supporting knowledge discovery application running in cloud environment as well as a holistic approach to predict the application execution times. We use rough sets theory to determine a reduct and then compute the execution time prediction. The heuristic reduct algorithm is based on frequencies of attributes appeared in discernibility matrix. We also propose to add dynamic information about the performances of various knowledge discovery tools over specific data sources to the Cloud Computing Environment for supporting the prediction. This information can be added as additional metadata stored in Cloud environment. Experimental result validates our solution that rough sets provide a formal framework for the problem of application execution time prediction in Cloud environment.


Keywords: Distributed computing, cloud computing, knowledge discovery, rough set.
 
Received April 25 2012; accepted January 17, 2013
  

Full Text

 
Print E-mail

Identifying Product Features from Customer Reviews Using Hybrid Dependency Patterns

Khairullah Khan1, Baharum Baharudin1, and Aurangzeb Khan2
1Computer and Information Sciences Department, Universiti Teknologi PETRONAS, Malaysia
2Institute of Engineering and Computing Sciences, University of Science & Technology Bannu, Pakistan

 
Abstract: In this paper we have addressed the problem of automatic identification of product features from customer reviews. Costumers, retailors, and manufacturers are popularly using customer reviews on websites for product reputation and sales forecasting. Opinion Mining application have been potentially employed to summarize the huge collectionof customer reviews for decision making. In this paper we have proposed hybrid dependency patterns to extract product features from unstructured reviews. The proposed dependency patterns exploit lexical relations and opinion context to identify features. Based on empirical analysis we found that the proposed hybrid patterns provide comparatively more accurate results. The average precision and recall are significantly improved with hybrid patterns.


Keywords: Opinion mining, features extraction, syntactic relation,context dependency.
 
Received February 3, 2012; accepted January 22, 2012
  

Full Text

 
Print E-mail

Volatility Modelling and Prediction by Hybrid Support Vector Regression with Chaotic Genetic Algorithms

Phichhang Ou and Hengshan Wang
Business School, University of Shanghai for Science and Technology, China

 
Abstract: In this paper, a new econometric model of volatility is proposed using hybrid Support Vector machine for Regression (SVR) combined with Chaotic Genetic Algorithm (CGA) to fit conditional mean and then conditional variance of stock market returns. The CGA, integrated by chaotic optimization algorithm (COA) with Genetic Algorithm (GA), is used to overcome premature local optimum in determining three hyperparameters of SVR model. The proposed hybrid SVRCGA model is achieved which includes the selection of input variables by ARMA approach for fitting both mean and variance functions of returns, and also the searching process of obtaining the optimal SVR hyperparameters based on the CGA while training the SVR. Real data of complex stock markets (NASDAQ) are applied to validate and check the predicting accuracy of the hybrid SVRCGA model. The experimental results showed that the proposed model outperforms the other competing models including SVR with GA, standard SVR, Kernel smoothing and several parametric GARCH type models.


Keywords: Chaotic optimization, GA, CGA, support vector regression, volatility.
 
Received November 15, 2011; accepted January 28, 2013
  

Full Text

 
Print E-mail

Comparative Performance Study of Several Features for Voiced/Non-Voiced Classification

Ykhlef Faycal1 and Bensebti Messaoud2
1Multimedia Laboratory, Centre de Développement des Technologies Avancées, Algeria
2Electronics Department, University of Saad Dahlab, Algeria

 
Abstract: This paper presents a comparative performance study of several time domain features for voiced/non-voiced classification of speech. Five classification schemes have been developed by combining one or two features amongst: Energy (E), Zeros Crossing Rate (ZCR), Autocorrelation Function (ACF), Average Magnitude Difference Function (AMDF), Weighted ACF (WACF), and the Discrete Wavelet Transform (DWT).   The development of these classifiers was based on the selection of the lowest number of time domain features which allow voicing decision without the need of any frequency transformation or pre processing approaches. The performance of the classifiers has been evaluated on speech data extracted from the TIMIT database. Two different noise types: white and babble, taken from the NOISEX92 database have been incorporated to validate the developed classification schemes in noisy environments. An overall ranking of these classifiers for high and low Signal to Noise Ratios (SNRs) have been established based on the average value of the Percentage of classification accuracy (Pc).


Keywords: ACF, AMDF, DWT, E, non-voiced, voiced, ZCR, WACF.
 
Received November 16, 2011; accepted January 28, 2013
  

Full Text

 
Print E-mail

An Efficient Parameters Selection for Object Recognition Based Colour Features in Traffic Image Retrieval

Hui Hui Wang1, Dzulkifli Mohamad2, Nor Azman Ismail2
1Faculty of Computer Science and Information Technology, Universiti Malaysia Sarawak, Malaysia
2Faculty of Computer Science and Information Systems, Universiti Teknologi Malaysia, Malaysia

 
Abstract: This paper proposes a novel technique for object identification and representation in complex traffic scene based on the colour features integrated with line detection techniques. Objects of interest (vehicles) are represented by using a Minimum Bound Region (MBR) with a reference coordinate. Object appearance is represented by colour-based features computed from the proposed technique. The performance of the object identification based colour features depends on some parameters which should be determined carefully to locate and identify objects that exists in the images successfully. Experiments have been conducted to determine the efficient parameters that should used and demonstrate that single and multiple known objects in complex scenes can be identified by the proposed approach.


Keywords: Colour features, Content based Image Retrieval (CBIR), Colour Object Recognition.
 
Received October 27, 2011; accepted May 22, 2012
  

Full Text

 
<
 
Copyright 2006-2009 Zarqa Private University. All rights reserved.
Print ISSN: 1683-3198.
 
 
Warning: fsockopen(): php_network_getaddresses: getaddrinfo failed: Name or service not known in /hsphere/local/home/ccis2k/ccis2k.org/iajit/templates/rt_chromatophore/index.php on line 251 Warning: fsockopen(): unable to connect to oucha.net:80 (php_network_getaddresses: getaddrinfo failed: Name or service not known) in /hsphere/local/home/ccis2k/ccis2k.org/iajit/templates/rt_chromatophore/index.php on line 251 skterr