Dimitris Varsamis | Technological Educational Institute (TEI) of Serres (original) (raw)

Papers by Dimitris Varsamis

Abstract: A new basis of interpolation points for the special case of the Newton two variable pol... more Abstract: A new basis of interpolation points for the special case of the Newton two variable polynomial interpolation problem is proposed. This basis is implemented when the upper bound of the total degree and the degree in each variable is known. It is shown that this new basis under certain conditions (that depends on the degrees of the interpolation polynomial), coincides either with the known triangular/rectangular basis or it is a polygonal basis. In all cases it uses the least interpolation points with further consequences to the complexity of the algorithms that we use.

The RPROP algorithm was originally developed in A fast and efficient training method for block-di... more The RPROP algorithm was originally developed in A fast and efficient training method for block-diagonal recurrent fuzzy neural networks is proposed. The method modifies the Simulated Annealing RPROP algorithm, originally developed for static models, in order to be applied to dynamic systems. A comparative analysis with a series of algorithms and recurrent models is given, indicating the effectiveness of the proposed learning approach. KEY WORDS In this context, this work proposes a modification of the standard SARPROP method, entitled Modified Simulated Annealing Resilient Back-propagation (M-SARPROP), which can be applied to a BDRNN, by taking into consideration the temporal relations existing in such a system. The rest of this paper is organized as follows: In Section 2 the structure and characteristics of the BDRNN are illustrated. The learning algorithm is developed in Section 3. In the next section a simulation example is presented, in order to highlight the behaviour of M-SARP...

World Academy of Science, Engineering and Technology, International Journal of Computer and Information Engineering, 2017

The aim of this work is the parallel implementation of k-means in MATLAB, in order to reduce the ... more The aim of this work is the parallel implementation of k-means in MATLAB, in order to reduce the execution time. Specifically, a new function in MATLAB for serial k-means algorithm is developed, which meets all the requirements for the conversion to a function in MATLAB with parallel computations. Additionally, two different variants for the definition of initial values are presented. In the sequel, the parallel approach is presented. Finally, the performance tests for the computation times respect to the numbers of features and classes are illustrated. Keywords—K-means algorithm, clustering, parallel computations, MATLAB.

In this work, we present the performance of three different parallel computing approaches of the ... more In this work, we present the performance of three different parallel computing approaches of the MATLAB Parallel Computing Toolbox. In particular, we use the command “parfor”, the command “spmd” and the technique “scheduler”. The comparison of the three approaches in terms of computations and memory are presented. The three approaches are applied to two specific problems: a) searching of a value into a matrix and b) prime factorization. The first problem is bounded by MATLAB for the size of matrix, namely, has memory problems, and the second problem is bounded by MATLAB for numerical precision and time complexity. Finally, the executions of the corresponding parallel algorithms in a multi-worker lab are presented.

Contemporary engineering sciences, 2020

In this article a novel parallel function of k-means algorithm is introduced reduceing significan... more In this article a novel parallel function of k-means algorithm is introduced reduceing significantly the computation time, compared to the serial equivalent of k-means in Matlab. A data set generator of hypergeometrical shapes clusters is additionally implemented for the evaluation of k-means algorithms in different datasets. The implemented parallel variations of k-means algorithm are functional at two alternative states, the state(k-meansRCP) where the centroids are assigned randomly and the state (k-meansRAP) where the data are assigned randomly to k clusters. The performance of the introduced k-meansRCP and k-meansRAP were compared to the serial implementations of k-means, on synthetic datasets of varying dimensionality.

WSEAS Transactions on Computers archive, 2018

In this work we introduce a parallel approach of Best Fit Decreasing algorithm. The Best Fit Decr... more In this work we introduce a parallel approach of Best Fit Decreasing algorithm. The Best Fit Decreasing algorithm is heuristic and is used for optimal assignment problems, for example cutting stock problem, bin packing problem etc. The above problems for optimal assignment have very large computational complexity. For this reason have developed heuristic algorithms which aim at the reduction of computational time with disadvantage on solution. The Best Fit Decreasing compute, in most times, a approach of optimal solution. The purpose of the study is twofold: (a) to split the dataset of problem with representative manner so that at every sub-problem to Best Fit Decreasing algorithm is applied and the cost to the results to be the smallest and (b) to be implemented program in Matlab that will running every sub-problem with parallel techniques with the aim of reducing computational time.

In this work a fast and efficient training method for block-diagonal recurrent neural networks is... more In this work a fast and efficient training method for block-diagonal recurrent neural networks is proposed. The method modifies and extends the Simulated Annealing RPROP algorithm, originally developed for static models, by taking into consideration the architectural characteristics and the temporal nature of this category of recurrent neural models. The performance of the proposed algorithm is evaluated through a comparative analysis with a series of algorithms and recurrent models.

This paper presents a computational intelligence-based filter that separates the adventitious dis... more This paper presents a computational intelligence-based filter that separates the adventitious discontinuous lung sounds from the vesicular sounds. The filter uses two Dynamic Fuzzy Neural Networks, operating in parallel, to perform the task of separation of the lung sounds, obtained from patients with pulmonary pathology. The Simulated Annealing Dynamic Resilient Propagation algorithm is employed for training the neurofuzzy system, and the resulting filter is applied to three major classes of lung sounds. The learning characteristics as well as the filter's separation qualities are highlighted by extensive experimental analysis and performance comparison with a series of other models.

Contemporary Engineering Sciences

In this work a parallel implementation of Best Fit Decreasing algorithm in Matlab is presented. T... more In this work a parallel implementation of Best Fit Decreasing algorithm in Matlab is presented. The propose of this work is twofold: (a) the reduction of the execution time and (b) the optimal partition of dataset for the minimum cost in results of algorithm. Specifically, a function for the partition of dataset is presented. Additionally, a function for BFD algorithm is developed, which meets all the requirements for parallel computations in Matlab. Finally, the performance tests for the computation times and the results of BFD algorithm respect to the number of dataset are illustrated.

Contemporary Engineering Sciences

In this paper a computational intelligence-based filter for real-time separation the adventitious... more In this paper a computational intelligence-based filter for real-time separation the adventitious discontinuous lung sounds from the vesicular sounds is proposed. The filter uses two Dynamic Fuzzy Neural Networks to perform the task of separation of the lung sounds, obtained from patients with pulmonary pathology. The networks are trained by the Simulated Annealing Dynamic Resilient Propagation algorithm and the resulting filter is applied to three major classes of lung sounds. In order to highlight the learning characteristics and the performance of the proposed separation scheme, extensive experimental analysis is conducted, where a comparison with other filters is given.

Applied Mathematical Sciences

In this paper a new parallel algorithm for the computation of the inverse of a bivariate polynomi... more In this paper a new parallel algorithm for the computation of the inverse of a bivariate polynomial matrix are presented. The parallel algorithm based on the technique evaluation-interpolation and for the part of interpolation uses the Newton bivariate polynomial interpolation. The algorithm is applied to the programming environment of MATLAB with Parallel Computing Toolbox and is compared to the corresponding build-in function of MATLAB inv().

In this work a fast and efficient training method for block-diagonal recurrent neural networks is... more In this work a fast and efficient training method for block-diagonal recurrent neural networks is proposed. The method modifies and extends the Simulated Annealing RPROP algorithm, originally developed for static models, by taking into consideration the architectural characteristics and the temporal nature of this category of recurrent neural models. The performance of the proposed algorithm is evaluated through a comparative analysis with a series of algorithms and recurrent models.

In this work a fast and numerical computational method for the calculation of determinant of a po... more In this work a fast and numerical computational method for the calculation of determinant of a polynomial matrix is proposed. The method modifies the Evaluation-Interpolation technique for the calculation of determinant and reduces the number of fixed required points to half with the use of complex basis. The performance of the proposed numerical computational method is evaluated through a comparative analysis with the simple computational method and built-in function of Matlab in software Matlab.

The purpose of this paper is to present an algorithmic formulation addressing the cartographic pr... more The purpose of this paper is to present an algorithmic formulation addressing the cartographic problem of siting an inset map at specific map locations under spatial and cartographic constraints. The first part of the paper aims at: (a) presenting a numerical algorithm that solves the above siting problem under such constraints, and (b) investigating the effectiveness of this numerical algorithm for a more general geographical problem, that of siting an anthropogenic structure or object of rectangular shape in suitable areas. The second part of this paper showcases the computational implementation of the above algorithm for addressing the cartographic problem of inset map placement in areas with land discontinuity.

5th Jubilee International Conference on Cartography & GIS

The aim of this work is twofold: a) the implementation of a computational procedure, where the ca... more The aim of this work is twofold: a) the implementation of a computational procedure, where the cartographic rules are defined as mathematical representations, and the geographic datasets are transformed into arithmetic data structures, b) the improvement of an existing searching algorithm to a more efficient one, with explicit use of the arithmetic data structures derived from the geographic datasets. This algorithm is implemented by the use of cartographic rules and is a part of a computational procedure which is applied into the Inset Mapper software tool, in order to help cartographers to tackle with the land discontinuity problem encountered in island cartography.

"Η Χαρτογραφία συχνά καλείται να αντιμετωπίσει προβλήματα για τα οποία η ανάπτυξη ειδικά σχεδιασμ... more "Η Χαρτογραφία συχνά καλείται να αντιμετωπίσει προβλήματα για τα οποία η ανάπτυξη ειδικά σχεδιασμένου λογισμικού κρίνεται απαραίτητη. Πιο συγκεκριμένα η Νησιωτική Χαρτογραφία αντιμετωπίζει ιδιαίτερα προβλήματα τα οποία εμφανίζονται κατά κύριο λόγο στην προσπάθεια απεικόνισης νησιωτικών περιοχών. Το πρόβλημα της «Χωρικής Ασυνέχειας του Γεωγραφικού Χώρου» αποτελεί ένα από τα πιο συχνά προβλήματα και εμφανίζεται κατά την οπτικοποίηση νησιωτικών περιοχών. Στην περίπτωση αυτή η κατασκευή ενός ενθέτου χάρτη αποτελεί την μοναδική και πιο ενδεδειγμένη χαρτογραφική λύση. Στο άρθρο αυτό αναλύεται το πρόβλημα της «Χωρικής Ασυνέχειας του Γεωγραφικού Χώρου», παρουσιάζονται ο τρόπος επίλυσης του με τη χρήση ένθετων χαρτών, καθώς και το θεωρητικό υπόβαθρο και η ομαδοποίηση των χαρτογραφικών κανόνων που διέπουν την κατασκευή ένθετων στη νησιωτική χαρτογραφία όπως ενσωματώνονται στην εφαρμογή Inset Mapper (IM). Επίσης, στο παρών άρθρο παρουσιάζεται μια ολοκληρωμένη περιγραφή όλων των λειτουργιών του IM, καθώς και τα πλεονεκτήματα που αποκομίζουν οι χαρτογράφοι από την χρήση του.

Λέξεις Κλειδιά: Νησιωτική Χαρτογραφία, Ένθετος Χάρτης, Ζητήματα Κλίμακας, Γεωοπτικοποίηση, Υπολογιστική Διαδικασία
"

FedCSIS 2012

""In this paper we present the implementation of a parallel searching algorithm, which is used f... more ""In this paper we present the implementation of a
parallel searching algorithm, which is used for the insetting
procedure in cartography. The calculation time of the above
procedure is very long due to the fact that the datasets in
cartography are maps with large and very large resolution. The
purpose of this proposal is to reduce the calculation time in a
multicore machine with shared memory. The proposed algorithm
and the performance tests are developed in Matlab Parallel
Toolbox.""

Applied Mathematics & Information Sciences, 2014

A new basis of interpolation points for the special case of the Newton two variable polynomial in... more A new basis of interpolation points for the special case of the Newton two variable polynomial interpolation problem is proposed. This basis is implemented when the upper bound of the total degree and the degree in each variable is known. It is shown that this new basis under certain conditions (that depends on the degrees of the interpolation polynomial), coincides either with the known triangular/rectangular basis or it is a polygonal basis. In all cases it uses the least interpolation points with further consequences to the complexity of the algorithms that we use.

2012 IEEE International Conference on Fuzzy Systems, 2012

ABSTRACT An application of fuzzy modeling to the problem of telecommunications data prediction is... more ABSTRACT An application of fuzzy modeling to the problem of telecommunications data prediction is proposed in this paper. The model building process is a two-stage sequential algorithm, based on the Orthogonal Least Squares (OLS) technique. Particularly, the OLS is first employed to partition the input space and determine the number of fuzzy rules and the premise parameters. In the sequel, a second orthogonal estimator determines the input terms which should be included in the consequent part of each fuzzy rule and calculate their parameters. Input selection is automatically performed, given a large input candidate set. Real world telecommunications data are used in order to highlight the characteristics of the proposed forecaster and to provide a comparative analysis with well-established forecasting models.

Applied Mathematics & Information Sciences, 2013

A recurrent neural network-based forecasting system for telecommunications call volume is propose... more A recurrent neural network-based forecasting system for telecommunications call volume is proposed in this work. In particular, the forecaster is a Block-Diagonal Recurrent Neural Network with internal feedback. Model's performance is evaluated by use of real-world telecommunications data, where an extensive comparative analysis with a series of existing forecasters is conducted, including both traditional models as well as neural and fuzzy approaches.

Abstract: A new basis of interpolation points for the special case of the Newton two variable pol... more Abstract: A new basis of interpolation points for the special case of the Newton two variable polynomial interpolation problem is proposed. This basis is implemented when the upper bound of the total degree and the degree in each variable is known. It is shown that this new basis under certain conditions (that depends on the degrees of the interpolation polynomial), coincides either with the known triangular/rectangular basis or it is a polygonal basis. In all cases it uses the least interpolation points with further consequences to the complexity of the algorithms that we use.

The RPROP algorithm was originally developed in A fast and efficient training method for block-di... more The RPROP algorithm was originally developed in A fast and efficient training method for block-diagonal recurrent fuzzy neural networks is proposed. The method modifies the Simulated Annealing RPROP algorithm, originally developed for static models, in order to be applied to dynamic systems. A comparative analysis with a series of algorithms and recurrent models is given, indicating the effectiveness of the proposed learning approach. KEY WORDS In this context, this work proposes a modification of the standard SARPROP method, entitled Modified Simulated Annealing Resilient Back-propagation (M-SARPROP), which can be applied to a BDRNN, by taking into consideration the temporal relations existing in such a system. The rest of this paper is organized as follows: In Section 2 the structure and characteristics of the BDRNN are illustrated. The learning algorithm is developed in Section 3. In the next section a simulation example is presented, in order to highlight the behaviour of M-SARP...

World Academy of Science, Engineering and Technology, International Journal of Computer and Information Engineering, 2017

The aim of this work is the parallel implementation of k-means in MATLAB, in order to reduce the ... more The aim of this work is the parallel implementation of k-means in MATLAB, in order to reduce the execution time. Specifically, a new function in MATLAB for serial k-means algorithm is developed, which meets all the requirements for the conversion to a function in MATLAB with parallel computations. Additionally, two different variants for the definition of initial values are presented. In the sequel, the parallel approach is presented. Finally, the performance tests for the computation times respect to the numbers of features and classes are illustrated. Keywords—K-means algorithm, clustering, parallel computations, MATLAB.

In this work, we present the performance of three different parallel computing approaches of the ... more In this work, we present the performance of three different parallel computing approaches of the MATLAB Parallel Computing Toolbox. In particular, we use the command “parfor”, the command “spmd” and the technique “scheduler”. The comparison of the three approaches in terms of computations and memory are presented. The three approaches are applied to two specific problems: a) searching of a value into a matrix and b) prime factorization. The first problem is bounded by MATLAB for the size of matrix, namely, has memory problems, and the second problem is bounded by MATLAB for numerical precision and time complexity. Finally, the executions of the corresponding parallel algorithms in a multi-worker lab are presented.

Contemporary engineering sciences, 2020

In this article a novel parallel function of k-means algorithm is introduced reduceing significan... more In this article a novel parallel function of k-means algorithm is introduced reduceing significantly the computation time, compared to the serial equivalent of k-means in Matlab. A data set generator of hypergeometrical shapes clusters is additionally implemented for the evaluation of k-means algorithms in different datasets. The implemented parallel variations of k-means algorithm are functional at two alternative states, the state(k-meansRCP) where the centroids are assigned randomly and the state (k-meansRAP) where the data are assigned randomly to k clusters. The performance of the introduced k-meansRCP and k-meansRAP were compared to the serial implementations of k-means, on synthetic datasets of varying dimensionality.

WSEAS Transactions on Computers archive, 2018

In this work we introduce a parallel approach of Best Fit Decreasing algorithm. The Best Fit Decr... more In this work we introduce a parallel approach of Best Fit Decreasing algorithm. The Best Fit Decreasing algorithm is heuristic and is used for optimal assignment problems, for example cutting stock problem, bin packing problem etc. The above problems for optimal assignment have very large computational complexity. For this reason have developed heuristic algorithms which aim at the reduction of computational time with disadvantage on solution. The Best Fit Decreasing compute, in most times, a approach of optimal solution. The purpose of the study is twofold: (a) to split the dataset of problem with representative manner so that at every sub-problem to Best Fit Decreasing algorithm is applied and the cost to the results to be the smallest and (b) to be implemented program in Matlab that will running every sub-problem with parallel techniques with the aim of reducing computational time.

In this work a fast and efficient training method for block-diagonal recurrent neural networks is... more In this work a fast and efficient training method for block-diagonal recurrent neural networks is proposed. The method modifies and extends the Simulated Annealing RPROP algorithm, originally developed for static models, by taking into consideration the architectural characteristics and the temporal nature of this category of recurrent neural models. The performance of the proposed algorithm is evaluated through a comparative analysis with a series of algorithms and recurrent models.

This paper presents a computational intelligence-based filter that separates the adventitious dis... more This paper presents a computational intelligence-based filter that separates the adventitious discontinuous lung sounds from the vesicular sounds. The filter uses two Dynamic Fuzzy Neural Networks, operating in parallel, to perform the task of separation of the lung sounds, obtained from patients with pulmonary pathology. The Simulated Annealing Dynamic Resilient Propagation algorithm is employed for training the neurofuzzy system, and the resulting filter is applied to three major classes of lung sounds. The learning characteristics as well as the filter's separation qualities are highlighted by extensive experimental analysis and performance comparison with a series of other models.

Contemporary Engineering Sciences

In this work a parallel implementation of Best Fit Decreasing algorithm in Matlab is presented. T... more In this work a parallel implementation of Best Fit Decreasing algorithm in Matlab is presented. The propose of this work is twofold: (a) the reduction of the execution time and (b) the optimal partition of dataset for the minimum cost in results of algorithm. Specifically, a function for the partition of dataset is presented. Additionally, a function for BFD algorithm is developed, which meets all the requirements for parallel computations in Matlab. Finally, the performance tests for the computation times and the results of BFD algorithm respect to the number of dataset are illustrated.

Contemporary Engineering Sciences

In this paper a computational intelligence-based filter for real-time separation the adventitious... more In this paper a computational intelligence-based filter for real-time separation the adventitious discontinuous lung sounds from the vesicular sounds is proposed. The filter uses two Dynamic Fuzzy Neural Networks to perform the task of separation of the lung sounds, obtained from patients with pulmonary pathology. The networks are trained by the Simulated Annealing Dynamic Resilient Propagation algorithm and the resulting filter is applied to three major classes of lung sounds. In order to highlight the learning characteristics and the performance of the proposed separation scheme, extensive experimental analysis is conducted, where a comparison with other filters is given.

Applied Mathematical Sciences

In this paper a new parallel algorithm for the computation of the inverse of a bivariate polynomi... more In this paper a new parallel algorithm for the computation of the inverse of a bivariate polynomial matrix are presented. The parallel algorithm based on the technique evaluation-interpolation and for the part of interpolation uses the Newton bivariate polynomial interpolation. The algorithm is applied to the programming environment of MATLAB with Parallel Computing Toolbox and is compared to the corresponding build-in function of MATLAB inv().

In this work a fast and efficient training method for block-diagonal recurrent neural networks is... more In this work a fast and efficient training method for block-diagonal recurrent neural networks is proposed. The method modifies and extends the Simulated Annealing RPROP algorithm, originally developed for static models, by taking into consideration the architectural characteristics and the temporal nature of this category of recurrent neural models. The performance of the proposed algorithm is evaluated through a comparative analysis with a series of algorithms and recurrent models.

In this work a fast and numerical computational method for the calculation of determinant of a po... more In this work a fast and numerical computational method for the calculation of determinant of a polynomial matrix is proposed. The method modifies the Evaluation-Interpolation technique for the calculation of determinant and reduces the number of fixed required points to half with the use of complex basis. The performance of the proposed numerical computational method is evaluated through a comparative analysis with the simple computational method and built-in function of Matlab in software Matlab.

The purpose of this paper is to present an algorithmic formulation addressing the cartographic pr... more The purpose of this paper is to present an algorithmic formulation addressing the cartographic problem of siting an inset map at specific map locations under spatial and cartographic constraints. The first part of the paper aims at: (a) presenting a numerical algorithm that solves the above siting problem under such constraints, and (b) investigating the effectiveness of this numerical algorithm for a more general geographical problem, that of siting an anthropogenic structure or object of rectangular shape in suitable areas. The second part of this paper showcases the computational implementation of the above algorithm for addressing the cartographic problem of inset map placement in areas with land discontinuity.

5th Jubilee International Conference on Cartography & GIS

The aim of this work is twofold: a) the implementation of a computational procedure, where the ca... more The aim of this work is twofold: a) the implementation of a computational procedure, where the cartographic rules are defined as mathematical representations, and the geographic datasets are transformed into arithmetic data structures, b) the improvement of an existing searching algorithm to a more efficient one, with explicit use of the arithmetic data structures derived from the geographic datasets. This algorithm is implemented by the use of cartographic rules and is a part of a computational procedure which is applied into the Inset Mapper software tool, in order to help cartographers to tackle with the land discontinuity problem encountered in island cartography.

"Η Χαρτογραφία συχνά καλείται να αντιμετωπίσει προβλήματα για τα οποία η ανάπτυξη ειδικά σχεδιασμ... more "Η Χαρτογραφία συχνά καλείται να αντιμετωπίσει προβλήματα για τα οποία η ανάπτυξη ειδικά σχεδιασμένου λογισμικού κρίνεται απαραίτητη. Πιο συγκεκριμένα η Νησιωτική Χαρτογραφία αντιμετωπίζει ιδιαίτερα προβλήματα τα οποία εμφανίζονται κατά κύριο λόγο στην προσπάθεια απεικόνισης νησιωτικών περιοχών. Το πρόβλημα της «Χωρικής Ασυνέχειας του Γεωγραφικού Χώρου» αποτελεί ένα από τα πιο συχνά προβλήματα και εμφανίζεται κατά την οπτικοποίηση νησιωτικών περιοχών. Στην περίπτωση αυτή η κατασκευή ενός ενθέτου χάρτη αποτελεί την μοναδική και πιο ενδεδειγμένη χαρτογραφική λύση. Στο άρθρο αυτό αναλύεται το πρόβλημα της «Χωρικής Ασυνέχειας του Γεωγραφικού Χώρου», παρουσιάζονται ο τρόπος επίλυσης του με τη χρήση ένθετων χαρτών, καθώς και το θεωρητικό υπόβαθρο και η ομαδοποίηση των χαρτογραφικών κανόνων που διέπουν την κατασκευή ένθετων στη νησιωτική χαρτογραφία όπως ενσωματώνονται στην εφαρμογή Inset Mapper (IM). Επίσης, στο παρών άρθρο παρουσιάζεται μια ολοκληρωμένη περιγραφή όλων των λειτουργιών του IM, καθώς και τα πλεονεκτήματα που αποκομίζουν οι χαρτογράφοι από την χρήση του.

Λέξεις Κλειδιά: Νησιωτική Χαρτογραφία, Ένθετος Χάρτης, Ζητήματα Κλίμακας, Γεωοπτικοποίηση, Υπολογιστική Διαδικασία
"

FedCSIS 2012

""In this paper we present the implementation of a parallel searching algorithm, which is used f... more ""In this paper we present the implementation of a
parallel searching algorithm, which is used for the insetting
procedure in cartography. The calculation time of the above
procedure is very long due to the fact that the datasets in
cartography are maps with large and very large resolution. The
purpose of this proposal is to reduce the calculation time in a
multicore machine with shared memory. The proposed algorithm
and the performance tests are developed in Matlab Parallel
Toolbox.""

Applied Mathematics & Information Sciences, 2014

A new basis of interpolation points for the special case of the Newton two variable polynomial in... more A new basis of interpolation points for the special case of the Newton two variable polynomial interpolation problem is proposed. This basis is implemented when the upper bound of the total degree and the degree in each variable is known. It is shown that this new basis under certain conditions (that depends on the degrees of the interpolation polynomial), coincides either with the known triangular/rectangular basis or it is a polygonal basis. In all cases it uses the least interpolation points with further consequences to the complexity of the algorithms that we use.

2012 IEEE International Conference on Fuzzy Systems, 2012

ABSTRACT An application of fuzzy modeling to the problem of telecommunications data prediction is... more ABSTRACT An application of fuzzy modeling to the problem of telecommunications data prediction is proposed in this paper. The model building process is a two-stage sequential algorithm, based on the Orthogonal Least Squares (OLS) technique. Particularly, the OLS is first employed to partition the input space and determine the number of fuzzy rules and the premise parameters. In the sequel, a second orthogonal estimator determines the input terms which should be included in the consequent part of each fuzzy rule and calculate their parameters. Input selection is automatically performed, given a large input candidate set. Real world telecommunications data are used in order to highlight the characteristics of the proposed forecaster and to provide a comparative analysis with well-established forecasting models.

Applied Mathematics & Information Sciences, 2013

A recurrent neural network-based forecasting system for telecommunications call volume is propose... more A recurrent neural network-based forecasting system for telecommunications call volume is proposed in this work. In particular, the forecaster is a Block-Diagonal Recurrent Neural Network with internal feedback. Model's performance is evaluated by use of real-world telecommunications data, where an extensive comparative analysis with a series of existing forecasters is conducted, including both traditional models as well as neural and fuzzy approaches.