Ecole Normale Superieure de Constantine

Constantine, Algeria

Ecole Normale Superieure de Constantine

Constantine, Algeria
SEARCH FILTERS
Time filter
Source Type

Hebboul A.,Ecole Normale Superieure de Constantine | Hachouf F.,University of Mentouri Constantine | Boulemnadjel A.,University of Mentouri Constantine
Neurocomputing | Year: 2015

In this paper, an Incremental Neural Network for Classification and Clustering (INNCC) is proposed. The main advantages of this neural network are the linkage between data topology preservation and classes representation by using the cluster posterior probabilities of classes. It is a constructive model without prior conditions such as a suitable number of nodes. A new neuron is inserted when new data are not represented by existing neurons. In training step, both supervised and unsupervised learning are used. The training dataset contains few samples with class labels and several unlabeled ones. The Support Vector Machines (SVM) operates in the training step to assess the INNCC classification result. The proposed approach has been tested on synthetic and real datasets. Obtained results are very promising. © 2015 Elsevier B.V.


Boulemnadjel A.,University of Mentouri Constantine | Hachouf F.,University of Mentouri Constantine | Hebboul A.,Ecole Normale Superieure de Constantine | Djemal K.,University of Évry Val d'Essonne
Engineering Applications of Artificial Intelligence | Year: 2015

In this paper a new soft subspace clustering algorithm is proposed. It is an iterative algorithm based on the minimization of a new objective function. The classification approach is developed by acting at three essential points. The first one is related to an initialization step; we suggest to use a multi-class support vector machine (SVM) for improving the initial classification parameters. The second point is based on the new objective function. It is formed by a separation term and compactness ones. The density of clusters is introduced in the last term to yield different cluster shapes. The third and the most important point consists in an active learning with SVM incorporated in the classification process. It allows a good estimation of the centers and the membership degrees and a speed convergence of the proposed algorithm. The developed approach has been tested to classify different synthetic datasets and real images databases. Several indices of performance have been used to demonstrate the superiority of the proposed method. Experimental results have corroborated the effectiveness of the proposed method in terms of good quality and optimized runtime. © 2015 Elsevier Ltd.


Boubekri M.,Ecole Normale Superieure de Constantine | Boubekri M.,University of Mentouri Constantine | Chaker A.,University of Mentouri Constantine
Energy Procedia | Year: 2011

In this work we propose to improve the production of active solar still by the use on the one hand, of external and internal reflectors in order to increase the rate of received overall solar radiation and on the other hand by using a thermal storage tank of in order to feed the still of the hot water for the night period. The input parameters of the computer program include the climatic conditions concerning Constantine town, Algeria (36 0 70′ N, 6 0 37′ E) for typical winter, spring and summer days. The results obtained show without ambiguity that the effect of the reflectors on the increase in the daily productivity of distillate is very significant for the winter period compared to the summer and spring. This increase is about 72.8 % in the winter. In addition it appears clearly that the night production of the still increases when this last is coupled with a storage tank out of the sunshine hours. This increase is about 27.54%, 21 % and 23.28% respectively for winter, spring and summer. © 2010 Published by Elsevier Ltd.


Hebboul A.,Ecole Normale Superieure de Constantine | Hachouf F.,University of Mentouri Constantine | Boulemnadjel A.,University of Mentouri Constantine
12th International Symposium on Programming and Systems, ISPS 2015 | Year: 2015

In this paper, we propose to reinforce the Self-Training strategy in a semi-supervised learning by using a Growing Probabilistic Neural Network (GPNN) which combines clustering and classification. The main advantages of this neural network are the linkage between data topology preservation and classes representation by using the cluster posterior probabilities of classes. It is a constructive model without prior conditions such as a suitable number of neurons. A new neuron is inserted when new data are not represented by existing neurons. For the Self-Training strategy, we chose the Support Vector Machines (SVM) as classifier because the SVMs are a powerful machine learning technique based on the principle of structural risk minimization. The proposed approach has been tested on synthetic and real datasets. Obtained results are very promising. © 2015 IEEE.


Zhu X.,University of Tennessee at Knoxville | Zhu X.,University of Chinese Academy of Sciences | Boushaba M.,University of Mentouri Constantine | Reghioua M.,Ecole Normale Superieure de Constantine
IEEE Transactions on Reliability | Year: 2015

The Joint Reliability Importance (JRI) of two components evaluates the interaction effect between the components on system reliability. This paper focuses on the JRI of components in a consecutive-K-out-of-n :F system, and an m-consecutive-k- out-of-n:F system, both with Markov-dependent components. We derive the closed-form formulas of the JRI of two components using probability generating functions, and extend the results to the JRI of three and more components. We further provide exact conditional distributions of random variables which are used in probability generating functions for determining the JRI of two components. Our numerical examples and tests demonstrate the use of derived formulas, and provide further insights about the JRI for Markov-dependent components. © 2015 IEEE.


Bellour A.,Ecole Normale Superieure de Constantine | Bousselsal M.,Laboratoire Dedp Non Lineaires Et Hm Ecole Normale Superieure Of Kouba
Mathematical Methods in the Applied Sciences | Year: 2014

This paper is concerned with the numerical solution of delay integro-differential equations. The main purpose of this work is to provide a new numerical approach based on the use of continuous collocation Taylor polynomials for the numerical solution of delay integro-differential equations. It is shown that this method is convergent. Numerical illustrations confirm our theoretical analysis. Copyright © 2013 John Wiley & Sons, Ltd.


Dehache I.,Ecole Normale Superieure de Constantine | Souici-Meslati L.,Annaba University
Proceedings of 2012 International Conference on Complex Systems, ICCS 2012 | Year: 2012

Nowadays, biometrics is a research field in full expansion, several identification and verification systems are now developed, however their performances remain unsatisfactory facing to the growing security needs. Generally, the use of only one biometric decreases the reliability of these systems; thus, we have to combine several modalities. In this paper, we propose a multibiometric fusion approach for identity verification using two modalities: the fingerprints and the signature. Combinations of neural multi-layer perceptrons (MLP) are used for the unimodal classification. Our multimodal integration approach is based on the use of Support Vector Machines (SVM). The final identity verification decision is made according to the scores generated by the SVM classifier. The experimental results of the proposed multibiometric system are encouraging. © 2012 IEEE.


Dehache I.,Ecole Normale Superieure de Constantine | Dehache I.,Annaba University | Souici-Meslati L.,Annaba University
Proceedings of 2015 IEEE World Conference on Complex Systems, WCCS 2015 | Year: 2015

Nowadays, breast cancer is very frequent among women. Early detection remains the only way to prevent this deadly disease and mammography is one of the most useful screening methods since the use of biopsy is unnecessary in most cases. In this paper, we propose a bio-inspired immunological approach for the classification of mammographic mass to distinguish malignant tumors from benign ones for computer-supported diagnosis. Our three classifiers are based on the artificial immune recognition algorithms AIRS1, AIRS2 and Parallel AIRS which represent three versions of the original Artificial Immune Recognition System AIRS. The obtained results are very promising and encourage the use of bio-inspired immunological approaches for medical data processing. © 2015 IEEE.


Boutamina S.,Ecole Normale Superieure de Constantine | Maamri R.,University Constantine2 Abdelhamid Mehri
ACM International Conference Proceeding Series | Year: 2015

This paper surveys recent research on context-aware workflow systems. It starts by addressing the notion of workflow and the different notions associated with it, for instance, the notion of "Process". Since the term "workflow" has different significations, several definitions are presented. In fact, Workflows are used in traditional computing environments and ubiquitous computing environments. In a ubiquitous computing environment, a workflow has to take into account the "context" which is generally represented as transition constraint for services execution. The adaptation of an application to changing circumstances and its response according to the context of use is known as Context-aware computing. Some context-aware workflow systems are presented to illustrate the representation of "context" in these systems. © 2015 ACM.


Haddad A.,Annaba University | Haddad A.,Ecole Normale Superieure de Constantine
Optoelectronics and Advanced Materials, Rapid Communications | Year: 2013

The determination and the understanding of the phenomena of elasticity in thin films are indispensable for the design and technology of various modern components. In this context, layer stiffening effects and mass loading effects are studied for several layer/substrate configurations via positive and negative dispersion curves respectively. The investigated structures showed four types of anomalies; a peak resulted of velocities greater than the Rayleigh velocity of the substrate, and a valley which is resulted of velocities smaller than the Rayleigh velocity of the layer, in the case of structures having 1 >VTL.VTS>1√2, as well as the inverse phenomenon, i.e. a valley comes from velocities smaller than the Rayleigh velocity of the substrate, and a peak obtained from velocities greater than the Rayleigh velocity of the layer, in the case of structures which having VTL/VTS>1. The anomalous behaviours appeared in the dispersion curves of Rayleigh velocity are studied in terms of the extreme velocity, VExt, which represents the maximum value of the peak, VMax, and the minimum value of the valley, VMin. The appearance and disappearance of any type of these phenomena is analysed and quantified in terms of a combined elastic parameter (δ). A general relationship which quantifies the extreme velocities in both cases, was deduced.

Loading Ecole Normale Superieure de Constantine collaborators
Loading Ecole Normale Superieure de Constantine collaborators