Search results for: Evolutionary algorithms

Authors: Randhir Singh Baghel

Abstract:

In this study, the fundamental ideas guiding the dynamic area of machine learning—where models thrive and algorithms change over time—are rooted in an innate mathematical link. This study explores the fundamental ideas that drive the development of intelligent systems, providing light on the mutually beneficial link between mathematics and machine learning.

Keywords: Machine Learning, deep learning, Neural Network, optimization.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 169

Authors: Stephen C. H. Leung, Defu Zhang

Abstract:

This paper addresses a stock-cutting problem with rotation of items and without the guillotine cutting constraint. In order to solve the large-scale problem effectively and efficiently, we propose a simple but fast heuristic algorithm. It is shown that this heuristic outperforms the latest published algorithms for large-scale problem instances.

Keywords: Combinatorial optimization, heuristic, large-scale, stock-cutting.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1683

Authors: A. R. Aswatha, T. Basavaraju, N. Bhaskara Rao

Abstract:

This paper presents a new technique for the optimum placement of processors to minimize the total effective communication load under multi-processor communication dominated environment. This is achieved by placing heavily loaded processors near each other and lightly loaded ones far away from one another in the physical grid locations. The results are mathematically proved for the Algorithms are described.

Keywords: Ascending Sort Index Vector, EffectiveCommunication Load, Effective Distance Matrix, OptimalPlacement, Sorting Order.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1349

Authors: Cecília Reis, J. A. Tenreiro Machado, J. Boaventura Cunha

Abstract:

This paper analyses the performance of a genetic algorithm using a new concept, namely a fractional-order dynamic fitness function, for the synthesis of combinational logic circuits. The experiments reveal superior results in terms of speed and convergence to achieve a solution.

Keywords: Circuit design, fractional-order systems, genetic algorithms, logic circuits

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1738

Authors: Kui Du

Abstract:

Recently, Uhlig [Numer. Algorithms, 52(3):335-353, 2009] proposed open questions about the ratios between the spectral norm, the numerical radius and the spectral radius of a square matrix. In this note, we provide some observations to answer these questions.

Keywords: Spectral norm, Numerical radius, Spectral radius, Ratios

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1826

Authors: Fahad Shahbaz Khan, Rao Muhammad Anwer, Olof Torgersson

Abstract:

The vast amount of information hidden in huge databases has created tremendous interests in the field of data mining. This paper examines the possibility of using data clustering techniques in oral medicine to identify functional relationships between different attributes and classification of similar patient examinations. Commonly used data clustering algorithms have been reviewed and as a result several interesting results have been gathered.

Keywords: Oral Medicine, Cluto, Data Clustering, Data Mining.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1977

Authors: Seung-Won Jung, Hye-Soo Kim, Le Thanh Ha, Seung-Jin Baek, Sung-Jea Ko

Abstract:

In this paper, a novel deinterlacing algorithm is proposed. The proposed algorithm approximates the distribution of the luminance into a polynomial function. Instead of using one polynomial function for all pixels, different polynomial functions are used for the uniform, texture, and directional edge regions. The function coefficients for each region are computed by matrix multiplications. Experimental results demonstrate that the proposed method performs better than the conventional algorithms.

Keywords: Deinterlacing, polynomial interpolation.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1382

Authors: Helle Hein, Ülo Lepik

Abstract:

A gradient learning method to regulate the trajectories of some nonlinear chaotic systems is proposed. The method is motivated by the gradient descent learning algorithms for neural networks. It is based on two systems: dynamic optimization system and system for finding sensitivities. Numerical results of several examples are presented, which convincingly illustrate the efficiency of the method.

Keywords: Chaos, Dynamical Systems, Learning, Neural Networks

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1367

Authors: Vadim Vagin, Marina Fomina, Oleg Morosin

Abstract:

This paper contains the description of argumentation approach for the problem of inductive concept formation. It is proposed to use argumentation, based on defeasible reasoning with justification degrees, to improve the quality of classification models, obtained by generalization algorithms. The experiment’s results on both clear and noisy data are also presented.

Keywords: Argumentation, justification degrees, inductive concept formation, noise, generalization.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1618

Authors: Amit Kr. Happy

Abstract:

This paper is motivated by the importance of multi-sensor image fusion with specific focus on Infrared (IR) and Visible image (VI) fusion for various applications including military reconnaissance. Image fusion can be defined as the process of combining two or more source images into a single composite image with extended information content that improves visual perception or feature extraction. These images can be from different modalities like Visible camera & IR Thermal Imager. While visible images are captured by reflected radiations in the visible spectrum, the thermal images are formed from thermal radiation (IR) that may be reflected or self-emitted. A digital color camera captures the visible source image and a thermal IR camera acquires the thermal source image. In this paper, some image fusion algorithms based upon Multi-Scale Transform (MST) and region-based selection rule with consistency verification have been proposed and presented. This research includes implementation of the proposed image fusion algorithm in MATLAB along with a comparative analysis to decide the optimum number of levels for MST and the coefficient fusion rule. The results are presented, and several commonly used evaluation metrics are used to assess the suggested method's validity. Experiments show that the proposed approach is capable of producing good fusion results. While deploying our image fusion algorithm approaches, we observe several challenges from the popular image fusion methods. While high computational cost and complex processing steps of image fusion algorithms provide accurate fused results, but they also make it hard to become deployed in system and applications that require real-time operation, high flexibility and low computation ability. So, the methods presented in this paper offer good results with minimum time complexity.

Keywords: Image fusion, IR thermal imager, multi-sensor, Multi-Scale Transform.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 430

Authors: Awadhesh Kumar Singh

Abstract:

Jayanti-s algorithm is one of the best known abortable mutual exclusion algorithms. This work is an attempt to overcome an already known limitation of the algorithm while preserving its all important properties and elegance. The limitation is that the token number used to assign process identification number to new incoming processes is unbounded. We have used a suitably adapted alternative data structure, in order to completely eliminate the use of token number, in the algorithm.

Keywords: Abortable, deterministic, local spin, mutual exclusion.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1273

Authors: Sergey Sorokin, Irina Sorokina, Alexander Yazenin

Abstract:

We present a solution to the Maxmin u/E parameters estimation problem of possibility distributions in m-dimensional case. Our method is based on geometrical approach, where minimal area enclosing ellipsoid is constructed around the sample. Also we demonstrate that one can improve results of well-known algorithms in fuzzy model identification task using Maxmin u/E parameters estimation.

Keywords: Possibility distribution, parameters estimation, Maxmin u/E estimator, fuzzy model identification.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2426

Authors: Christian H. Sanabria-Montaña, Rodrigo Huerta-Quintanilla

Abstract:

A lattice network is a special type of network in which all nodes have the same number of links, and its boundary conditions are periodic. The most basic lattice network is the ring, a one-dimensional network with periodic border conditions. In contrast, the Cartesian product of d rings forms a d-dimensional lattice network. An analytical expression currently exists for the clustering coefficient in this type of network, but the theoretical value is valid only up to certain connectivity value; in other words, the analytical expression is incomplete. Here we obtain analytically the clustering coefficient expression in d-dimensional lattice networks for any link density. Our analytical results show that the clustering coefficient for a lattice network with density of links that tend to 1, leads to the value of the clustering coefficient of a fully connected network. We developed a model on criminology in which the generalized clustering coefficient expression is applied. The model states that delinquents learn the know-how of crime business by sharing knowledge, directly or indirectly, with their friends of the gang. This generalization shed light on the network properties, which is important to develop new models in different fields where network structure plays an important role in the system dynamic, such as criminology, evolutionary game theory, econophysics, among others.

Keywords: Clustering coefficient, criminology, generalized, regular network d-dimensional.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1636

Authors: C. A. Barros, Ana P. Barroso

Abstract:

Any industrial company needs to determine the amount of variation that exists within its measurement process and guarantee the reliability of their data, studying the performance of their measurement system, in terms of linearity, bias, repeatability and reproducibility and stability. This issue is critical for automotive industry suppliers, who are required to be certified by the 16949:2016 standard (replaces the ISO/TS 16949) of International Automotive Task Force, defining the requirements of a quality management system for companies in the automotive industry. Measurement System Analysis (MSA) is one of the mandatory tools. Frequently, the measurement system in companies is not connected to the equipment and do not incorporate the methods proposed by the Automotive Industry Action Group (AIAG). To address these constraints, an R&D project is in progress, whose objective is to develop a web and cloud-based MSA tool. This MSA tool incorporates Industry 4.0 concepts, such as, Internet of Things (IoT) protocols to assure the connection with the measuring equipment, cloud computing, artificial intelligence, statistical tools, and advanced mathematical algorithms. This paper presents the preliminary findings of the project. The web and cloud-based MSA tool is innovative because it implements all statistical tests proposed in the MSA-4 reference manual from AIAG as well as other emerging methods and techniques. As it is integrated with the measuring devices, it reduces the manual input of data and therefore the errors. The tool ensures traceability of all performed tests and can be used in quality laboratories and in the production lines. Besides, it monitors MSAs over time, allowing both the analysis of deviations from the variation of the measurements performed and the management of measurement equipment and calibrations. To develop the MSA tool a ten-step approach was implemented. Firstly, it was performed a benchmarking analysis of the current competitors and commercial solutions linked to MSA, concerning Industry 4.0 paradigm. Next, an analysis of the size of the target market for the MSA tool was done. Afterwards, data flow and traceability requirements were analysed in order to implement an IoT data network that interconnects with the equipment, preferably via wireless. The MSA web solution was designed under UI/UX principles and an API in python language was developed to perform the algorithms and the statistical analysis. Continuous validation of the tool by companies is being performed to assure real time management of the ‘big data’. The main results of this R&D project are: MSA Tool, web and cloud-based; Python API; New Algorithms to the market; and Style Guide of UI/UX of the tool. The MSA tool proposed adds value to the state of the art as it ensures an effective response to the new challenges of measurement systems, which are increasingly critical in production processes. Although the automotive industry has triggered the development of this innovative MSA tool, other industries would also benefit from it. Currently, companies from molds and plastics, chemical and food industry are already validating it.

Keywords: Automotive industry, Industry 4.0, internet of things, IATF 16949:2016, measurement system analysis.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 993

Authors: S. Bougrine, S. Ouchraa, B. Ahiod, A. A. El Imrani

Abstract:

Ant colony optimization is an ant algorithm framework that took inspiration from foraging behavior of ant colonies. Indeed, ACO algorithms use a chemical communication, represented by pheromone trails, to build good solutions. However, ants involve different communication channels to interact. Thus, this paper introduces the acoustic communication between ants while they are foraging. This process allows fine and local exploration of search space and permits optimal solution to be improved.

Keywords: Acoustic Communication, Ant Colony Optimization, Local Search, Traveling Salesman Problem.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2433

Authors: Ali Habiboghli, Rouhollah Mostafaei, Vasif Nabiyev

Abstract:

Optical network uses a tool for routing called Latin router. These routers use particular algorithms for routing. For example, we can refer to LDF algorithm that uses backtracking (one of CSP methods) for problem solving. In this paper, we proposed new approached for completion routing table (DRA&CRA algorithm) and compare with pervious proposed ways and showed numbers of backtracking, blocking and run time for DRA algorithm less than LDF and CRA algorithm.

Keywords: Latin Router, Constraint Satisfaction Problem, Wavelength Routing.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1655

Authors: Okay Gunes

Abstract:

Unsatisfactory experiences due to an information shortage regarding the future pay-offs of actual choices, yield satisficing decision-making. This research will examine, for the first time in the literature, the motivation behind suboptimal decisions due to uncertainty by subjecting Adam Smith’s and Jeremy Bentham’s assumptions about the nature of the actions that lead to satisficing behavior, in order to clarify the theoretical background of a “consumption-based satisfactory time” concept. The contribution of this paper with respect to the existing literature is threefold: firstly, it is showed in this paper that Adam Smith’s uncertainty is related to the problem of the constancy of ideas and not related directly to beliefs. Secondly, possessions, as in Jeremy Bentham’s oeuvre, are assumed to be just as pleasing, as protecting and improving the actual or expected quality of life, so long as they reduce any displeasure due to the undesired outcomes of uncertainty. Finally, each consumption decision incurs its own satisfactory time period, owed to not feeling hungry, being healthy, not having transportation…etc. This reveals that the level of satisfaction is indeed a behavioral phenomenon where its value would depend on the simultaneous satisfaction derived from all activities.

Keywords: Decision-making, idea and belief, satisficing, uncertainty.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1930

Authors: Dragos Nicolae VIZIREANU

Abstract:

One of the main image representations in Mathematical Morphology is the 3D Shape Decomposition Representation, useful for Image Compression and Representation,and Pattern Recognition. The 3D Morphological Shape Decomposition representation can be generalized a number of times,to extend the scope of its algebraic characteristics as much as possible. With these generalizations, the Morphological Shape Decomposition 's role to serve as an efficient image decomposition tool is extended to grayscale images.This work follows the above line, and further develops it. Anew evolutionary branch is added to the 3D Morphological Shape Decomposition's development, by the introduction of a 3D Multi Structuring Element Morphological Shape Decomposition, which permits 3D Morphological Shape Decomposition of 3D binary images (grayscale images) into "multiparameter" families of elements. At the beginning, 3D Morphological Shape Decomposition representations are based only on "1 parameter" families of elements for image decomposition.This paper addresses the gray scale inter frame interpolation by means of mathematical morphology. The new interframe interpolation method is based on generalized morphological 3D Shape Decomposition. This article will present the theoretical background of the morphological interframe interpolation, deduce the new representation and show some application examples.Computer simulations could illustrate results.

Keywords: 3D shape decomposition representation, mathematical morphology, gray scale interframe interpolation

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1744

Authors: Narasimham Challa, Jayaram Pradhan

Abstract:

Cryptographic algorithms play a crucial role in the information society by providing protection from unauthorized access to sensitive data. It is clear that information technology will become increasingly pervasive, Hence we can expect the emergence of ubiquitous or pervasive computing, ambient intelligence. These new environments and applications will present new security challenges, and there is no doubt that cryptographic algorithms and protocols will form a part of the solution. The efficiency of a public key cryptosystem is mainly measured in computational overheads, key size and bandwidth. In particular the RSA algorithm is used in many applications for providing the security. Although the security of RSA is beyond doubt, the evolution in computing power has caused a growth in the necessary key length. The fact that most chips on smart cards can-t process key extending 1024 bit shows that there is need for alternative. NTRU is such an alternative and it is a collection of mathematical algorithm based on manipulating lists of very small integers and polynomials. This allows NTRU to high speeds with the use of minimal computing power. NTRU (Nth degree Truncated Polynomial Ring Unit) is the first secure public key cryptosystem not based on factorization or discrete logarithm problem. This means that given sufficient computational resources and time, an adversary, should not be able to break the key. The multi-party communication and requirement of optimal resource utilization necessitated the need for the present day demand of applications that need security enforcement technique .and can be enhanced with high-end computing. This has promoted us to develop high-performance NTRU schemes using approaches such as the use of high-end computing hardware. Peer-to-peer (P2P) or enterprise grids are proven as one of the approaches for developing high-end computing systems. By utilizing them one can improve the performance of NTRU through parallel execution. In this paper we propose and develop an application for NTRU using enterprise grid middleware called Alchemi. An analysis and comparison of its performance for various text files is presented.

Keywords: Alchemi, GridNtru, Ntru, PKCS.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1691

Authors: Pınar Tüfekçi, Yılmaz Kılıçaslan

Abstract:

In this paper we present a computational model for pronominal anaphora resolution in Turkish. The model is based on Hobbs’ Naїve Algorithm [4, 5, 6], which exploits only the surface syntax of sentences in a given text.

Keywords: Anaphora Resolution, Pronoun Resolution, Syntax based Algorithms, Naїve Algorithm.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 3238

Authors: Jiashang Jiang, Hao Liu, Yongxin Yuan

Abstract:

In this paper the gradient based iterative algorithms are presented to solve the following four types linear matrix equations: (a) AXB = F; (b) AXB = F, CXD = G; (c) AXB = F s. t. X = XT ; (d) AXB+CYD = F, where X and Y are unknown matrices, A,B,C,D, F,G are the given constant matrices. It is proved that if the equation considered has a solution, then the unique minimum norm solution can be obtained by choosing a special kind of initial matrices. The numerical results show that the proposed method is reliable and attractive.

Keywords: Matrix equation, iterative algorithm, parameter estimation, minimum norm solution.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1857

Authors: Tsuyoshi Okita, Tatsuya Niwa

Abstract:

For a given specific problem an efficient algorithm has been the matter of study. However, an alternative approach orthogonal to this approach comes out, which is called a reduction. In general for a given specific problem this reduction approach studies how to convert an original problem into subproblems. This paper proposes a formal modeling language to support this reduction approach. We show three examples from the wide area of learning problems. The benefit is a fast prototyping of algorithms for a given new problem.

Keywords: Formal language, statistical inference problem, reduction.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1614

Authors: Konstantinos Liagkouras, Konstantinos Metaxiotis

Abstract:

In this paper we present a classification of the various technologies applied for the solution of the portfolio selection problem according to the discipline and the methodological framework followed. We provide a concise presentation of the emerged categories and we are trying to identify which methods considered obsolete and which lie at the heart of the debate. On top of that, we provide a comparative study of the different technologies applied for efficient portfolio construction and we suggest potential paths for future work that lie at the intersection of the presented techniques.

Keywords: Portfolio selection, optimization techniques, financial models, stochastics, heuristics.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1751

Authors: O. Sotolongo-Grau, D. Rodriguez-Perez, J. A. Santos-Miranda, M. M. Desco, O. Sotolongo-Costa, J. C. Antoranz

Abstract:

The use of the oncologic index ISTER allows for a more effective planning of the radiotherapic facilities in the hospitals. Any change in the radiotherapy treatment, due to unexpected stops, may be adapted by recalculating the doses to the new treatment duration while keeping the optimal prognosis. The results obtained in a simulation model on millions of patients allow the definition of optimal success probability algorithms.

Keywords: Mathematical model, radiation oncology, dynamical systems applications.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1551

Authors: C. Chahin, M. C. Hsu, Y. H. Lin, C. Y. Huang

Abstract:

Electronics Products that achieve high levels of integrated communications, computing and entertainment, multimedia features in small, stylish and robust new form factors are winning in the market place. Due to the high costs that an industry may undergo and how a high yield is directly proportional to high profits, IC (Integrated Circuit) manufacturers struggle to maximize yield, but today-s customers demand miniaturization, low costs, high performance and excellent reliability making the yield maximization a never ending research of an enhanced assembly process. With factors such as minimum tolerances, tighter parameter variations a systematic approach is needed in order to predict the assembly process. In order to evaluate the quality of upcoming circuits, yield models are used which not only predict manufacturing costs but also provide vital information in order to ease the process of correction when the yields fall below expectations. For an IC manufacturer to obtain higher assembly yields all factors such as boards, placement, components, the material from which the components are made of and processes must be taken into consideration. Effective placement yield depends heavily on machine accuracy and the vision of the system which needs the ability to recognize the features on the board and component to place the device accurately on the pads and bumps of the PCB. There are currently two methods for accurate positioning, using the edge of the package and using solder ball locations also called footprints. The only assumption that a yield model makes is that all boards and devices are completely functional. This paper will focus on the Monte Carlo method which consists in a class of computational algorithms (information processed algorithms) which depends on repeated random samplings in order to compute the results. This method utilized in order to recreate the simulation of placement and assembly processes within a production line.

Keywords: Monte Carlo simulation, placement yield, PCBcharacterization, electronics assembly

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2166

Authors: Jae Wook Shin, Ju-man Song, Hyun-Taek Choi, Poo Gyeon Park

Abstract:

This paper introduces a new variable step-size APA with decorrelation of AR input process is based on the MSD analysis. To achieve a fast convergence rate and a small steady-state estimation error, he proposed algorithm uses variable step size that is determined by minimising the MSD. In addition, experimental results show that the proposed algorithm is achieved better performance than the other algorithms.

Keywords: adaptive filter, affine projection algorithm, variable step size.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1897

Authors: Satoshi Shimizu, Kazuaki Yamaguchi, Toshiki Saitoh, Sumio Masuda

Abstract:

Some fast exact algorithms for the maximum weight clique problem have been proposed. Östergard’s algorithm is one of them. Kumlander says his algorithm is faster than it. But we confirmed that the straightforwardly implemented Kumlander’s algorithm is slower than O¨ sterga˚rd’s algorithm. We propose some improvements on Kumlander’s algorithm.

Keywords: Maximum weight clique, exact algorithm, branch-andbound, NP-hard.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1858

Authors: Young-Seok Choi

Abstract:

We present a normalized LMS (NLMS) algorithm with robust regularization. Unlike conventional NLMS with the fixed regularization parameter, the proposed approach dynamically updates the regularization parameter. By exploiting a gradient descent direction, we derive a computationally efficient and robust update scheme for the regularization parameter. In simulation, we demonstrate the proposed algorithm outperforms conventional NLMS algorithms in terms of convergence rate and misadjustment error.

Keywords: Regularization, normalized LMS, system identification, robustness.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1876

Authors: Sheng-Gwo Chen, Wen-Haw Chen

Abstract:

It is an important problem to compute the geodesics on a surface in many fields. To find the geodesics in practice, however, the traditional discrete algorithms or numerical approaches can only find a list of discrete points. The first author proposed in 2010 a new, elegant and accurate method, the geodesic-like method, for approximating geodesics on a regular surface. This paper will present by use of this method a computation of the Bezier geodesic-like curves on spheres.

Keywords: Geodesics, Geodesic-like curve, Spheres, Bezier.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1622

Authors: Monjed Ezzeddinne, Nicolas Lefebvre, Régis Lengellé

Abstract:

This paper presents a new version of the SVM mixture algorithm initially proposed by Kwok for classification and regression problems. For both cases, a slight modification of the mixture model leads to a standard SVM training problem, to the existence of an exact solution and allows the direct use of well known decomposition and working set selection algorithms. Only the regression case is considered in this paper but classification has been addressed in a very similar way. This method has been successfully applied to engine pollutants emission modeling.

Keywords: Identification, Learning systems, Mixture ofExperts, Support Vector Machines.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1365