Category: Uncategorized

A hierarchical multiclass support vector machine incorporated with holistic triple learning units

Abstract  This paper proposes a new hierarchical learning structure, namely the holistic triple learning (HTL), for extending the binary
support vector machine (SVM) to multi-classification problems. For an N-class problem, a HTL constructs …

Abstract  

This paper proposes a new hierarchical learning structure, namely the holistic triple learning (HTL), for extending the binary
support vector machine (SVM) to multi-classification problems. For an N-class problem, a HTL constructs a decision tree up to a depth of

éN/3ù+1

. A leaf node of the decision tree is allowed to be placed with a holistic triple learning unit whose generalisation abilities
are assessed and approved. Meanwhile, the remaining nodes in the decision tree each accommodate a standard binary SVM classifier.
The holistic triple classifier is a regression model trained on three classes, whose training algorithm is originated from
a recently proposed implementation technique, namely the least-squares support vector machine (LS-SVM). A major novelty with
the holistic triple classifier is the reduced number of support vectors in the solution. For the resultant HTL-SVM, an upper
bound of the generalisation error can be obtained. The time complexity of training the HTL-SVM is analysed, and is shown to
be comparable to that of training the one-versus-one (1-vs.-1) SVM, particularly on small-scale datasets. Empirical studies
show that the proposed HTL-SVM achieves competitive classification accuracy with a reduced number of support vectors compared
to the popular 1-vs-1 alternative.

  • Content Type Journal Article
  • Pages 1-11
  • DOI 10.1007/s00500-010-0551-9
  • Authors
    • Xiao-Lei Xia, Queen’s University of Belfast Intelligent Systems and Control, School of Electronics, Electrical Engineering and Computer Science Belfast BT9 5AH UK
    • Kang Li, Queen’s University of Belfast Intelligent Systems and Control, School of Electronics, Electrical Engineering and Computer Science Belfast BT9 5AH UK
    • George W. Irwin, Queen’s University of Belfast Intelligent Systems and Control, School of Electronics, Electrical Engineering and Computer Science Belfast BT9 5AH UK

Recent progress in natural computation and knowledge discovery: an ICNC’09-FSKD’09 special issue

Recent progress in natural computation and knowledge discovery: an ICNC’09-FSKD’09 special issue
Content Type Journal ArticlePages 1-2DOI 10.1007/s00500-010-0559-1Authors
Haiying Wang, University of Ulster School of Computing and Mathematics and…

Recent progress in natural computation and knowledge discovery: an ICNC’09-FSKD’09 special issue

  • Content Type Journal Article
  • Pages 1-2
  • DOI 10.1007/s00500-010-0559-1
  • Authors
    • Haiying Wang, University of Ulster School of Computing and Mathematics and Computer Science Research Institute Shore Road Newtownabbey Co. Antrim BT37 0QB UK
    • Yixin Chen, Washington University in St Louis Computer Science and Engineering Campus Box 1045, One Brookings Drive St. Louis MO 63130 USA
    • Hepu Deng, RMIT University Business Information Technology GPO Box 2476 Melbourne VIC 3001 Australia
    • Lipo Wang, Nanyang Technological University School of Electrical and Electronic Engineering Block S1, 50 Nanyang Avenue Singapore 639798 Singapore

A novel approach to annotating web service based on interface concept mapping and semantic expansion

Abstract  With the rapid development of web service technology in these years, traditional standards have been matured during the process
of service registry and discovery. However, it is difficult for service requesters to discover satisfac…

Abstract  

With the rapid development of web service technology in these years, traditional standards have been matured during the process
of service registry and discovery. However, it is difficult for service requesters to discover satisfactory web services.
The reason for this phenomenon is that the traditional service organization mode lacks semantic understanding ability for
service function interface. This paper proposes a novel approach to annotating web services. We first adopt domain ontology
as a semantic context, and give our general framework of service semantic annotation. Then, interface concept mapping algorithm
and service interface expansion algorithm are respectively presented in detail. Finally, the generation process of semantic
web service repository is presented based on preceding algorithms. Simulation experiment results demonstrate that annotated
web services by the proposed method can more satisfy requirements for service requesters than traditional ones by service
matchmaking engine. It can get better service discovery effectiveness.

  • Content Type Journal Article
  • Pages 1-10
  • DOI 10.1007/s00500-010-0548-4
  • Authors
    • Guobing Zou, Tongji University Department of Computer Science and Technology Shanghai 201804 China
    • Yang Xiang, Tongji University Department of Computer Science and Technology Shanghai 201804 China
    • Yanglan Gan, Tongji University Department of Computer Science and Technology Shanghai 201804 China
    • Yixin Chen, Washington University Department of Computer Science and Engineering St. Louis MO 63130 USA

A novel multi-population cultural algorithm adopting knowledge migration

Abstract  In existing multi-population cultural algorithms, information is exchanged among sub-populations by individuals. However,
migrated individuals cannot reflect enough evolutionary information, which limits the evolution performance. …

Abstract  

In existing multi-population cultural algorithms, information is exchanged among sub-populations by individuals. However,
migrated individuals cannot reflect enough evolutionary information, which limits the evolution performance. In order to enhance
the migration efficiency, a novel multi-population cultural algorithm adopting knowledge migration is proposed. Implicit knowledge
extracted from the evolution process of each sub-population directly reflects the information about dominant search space.
By migrating knowledge among sub-populations at the constant intervals, the algorithm realizes more effective interaction
with less communication cost. Taken benchmark functions with high-dimension as the examples, simulation results indicate that
the algorithm can effectively improve the speed of convergence and overcome premature convergence.

  • Content Type Journal Article
  • Pages 1-9
  • DOI 10.1007/s00500-010-0556-4
  • Authors
    • Yi-nan Guo, College of Information and Electronic Engineering, China University of Mining and Technology, Xuzhou, 221008 Jiangsu China
    • Jian Cheng, College of Information and Electronic Engineering, China University of Mining and Technology, Xuzhou, 221008 Jiangsu China
    • Yuan-yuan Cao, College of Information and Electronic Engineering, China University of Mining and Technology, Xuzhou, 221008 Jiangsu China
    • Yong Lin, College of Information and Electronic Engineering, China University of Mining and Technology, Xuzhou, 221008 Jiangsu China

Foreword

Foreword
Content Type Journal ArticleDOI 10.1007/s11047-010-9181-5Authors
Friedrich Simmel, Technische Universität München Garching GermanyAshish Goel, Stanford University Stanford CA USA

Journal Natural ComputingOnline ISSN 1572-9796Print …

Foreword

  • Content Type Journal Article
  • DOI 10.1007/s11047-010-9181-5
  • Authors
    • Friedrich Simmel, Technische Universität München Garching Germany
    • Ashish Goel, Stanford University Stanford CA USA

Using selfish gene theory to construct mutual information and entropy based clusters for bivariate optimizations

Abstract  This paper proposes a new approach named SGMIEC in the field of estimation of distribution algorithm (EDA). While the current
EDAs require much time in the statistical learning process as the relationships among the variables are t…

Abstract  This paper proposes a new approach named SGMIEC in the field of estimation of distribution algorithm (EDA). While the current
EDAs require much time in the statistical learning process as the relationships among the variables are too complicated, the
selfish gene theory (SG) is deployed in this approach and a mutual information and entropy based cluster (MIEC) model with
an incremental learning and resample scheme is also set to optimize the probability distribution of the virtual population.
Experimental results on several benchmark problems demonstrate that, compared with BMDA, COMIT and MIMIC, SGMIEC often performs
better in convergent reliability, convergent velocity and convergent process.

  • Content Type Journal Article
  • Pages 1-9
  • DOI 10.1007/s00500-010-0557-3
  • Authors
    • Feng Wang, Wuhan University State Key Laboratory of Software Engineering Wuhan China
    • Zhiyi Lin, Wuhan University State Key Laboratory of Software Engineering Wuhan China
    • Cheng Yang, Wuhan University State Key Laboratory of Software Engineering Wuhan China
    • Yuanxiang Li, Wuhan University State Key Laboratory of Software Engineering Wuhan China

Environment identification-based memory scheme for estimation of distribution algorithms in dynamic environments

Abstract  In estimation of distribution algorithms (EDAs), the joint probability distribution of high-performance solutions is presented
by a probability model. This means that the priority search areas of the solution space are characterize…

Abstract  In estimation of distribution algorithms (EDAs), the joint probability distribution of high-performance solutions is presented
by a probability model. This means that the priority search areas of the solution space are characterized by the probability
model. From this point of view, an environment identification-based memory management scheme (EI-MMS) is proposed to adapt
binary-coded EDAs to solve dynamic optimization problems (DOPs). Within this scheme, the probability models that characterize
the search space of the changing environment are stored and retrieved to adapt EDAs according to environmental changes. A
diversity loss correction scheme and a boundary correction scheme are combined to counteract the diversity loss during the
static evolutionary process of each environment. Experimental results show the validity of the EI-MMS and indicate that the
EI-MMS can be applied to any binary-coded EDAs. In comparison with three state-of-the-art algorithms, the univariate marginal
distribution algorithm (UMDA) using the EI-MMS performs better when solving three decomposable DOPs. In order to understand
the EI-MMS more deeply, the sensitivity analysis of parameters is also carried out in this paper.

  • Content Type Journal Article
  • Pages 1-16
  • DOI 10.1007/s00500-010-0547-5
  • Authors
    • Xingguang Peng, Northwestern Polytechnical University School of Electronics and Information Xi’an Shaanxi 710129 China
    • Xiaoguang Gao, Northwestern Polytechnical University School of Electronics and Information Xi’an Shaanxi 710129 China
    • Shengxiang Yang, University of Leicester Department of Computer Science University Road Leicester LE1 7RH UK

Solution of fuzzy polynomial equations by modified Adomian decomposition method

Abstract  In this paper, we present some efficient numerical algorithm for solving fuzzy polynomial equations based on Newton’s method.
The modified Adomian decomposition method is applied to construct the numerical algorithms. Some numeri…

Abstract  In this paper, we present some efficient numerical algorithm for solving fuzzy polynomial equations based on Newton’s method.
The modified Adomian decomposition method is applied to construct the numerical algorithms. Some numerical illustrations are
given to show the efficiency of algorithms.

  • Content Type Journal Article
  • Pages 1-6
  • DOI 10.1007/s00500-010-0546-6
  • Authors
    • M. Otadi, Islamic Azad University Department of Mathematics Firuozkooh Branch Firuozkooh Iran
    • M. Mosleh, Islamic Azad University Department of Mathematics Firuozkooh Branch Firuozkooh Iran

Extremal states on bounded residuated -monoids with general comparability

Abstract  Bounded residuated lattice ordered monoids (

Rl
-monoids) are a common generalization of pseudo-

BL
-algebras and Heyting algebras, i.e. algebras of the non-commutative basic fuzzy logic (and consequently of the basic fuzzy
logic…

Abstract  Bounded residuated lattice ordered monoids (

Rl

-monoids) are a common generalization of pseudo-

BL

-algebras and Heyting algebras, i.e. algebras of the non-commutative basic fuzzy logic (and consequently of the basic fuzzy
logic, the Łukasiewicz logic and the non-commutative Łukasiewicz logic) and the intuitionistic logic, respectively. We investigate
bounded

Rl

-monoids satisfying the general comparability condition in connection with their states (analogues of probability measures).
It is shown that if an extremal state on Boolean elements fulfils a simple condition, then it can be uniquely extended to
an extremal state on the

Rl

-monoid, and that if every extremal state satisfies this condition, then the

Rl

-monoid is a pseudo-

BL

-algebra.

  • Content Type Journal Article
  • Pages 1-5
  • DOI 10.1007/s00500-010-0545-7
  • Authors
    • Jiří Rachůnek, Palacký University Department of Algebra and Geometry, Faculty of Sciences Tomkova 40 779 00 Olomouc Czech Republic
    • Dana Šalounová, VŠB-Technical University Ostrava Sokolská 33 701 21 Ostrava Czech Republic

Evolving robust GP solutions for hedge fund stock selection in emerging markets

Abstract  Stock selection for hedge fund portfolios is a challenging problem for Genetic Programming (GP) because the markets (the environment
in which the GP solution must survive) are dynamic, unpredictable and unforgiving. How can GP be i…

Abstract  Stock selection for hedge fund portfolios is a challenging problem for Genetic Programming (GP) because the markets (the environment
in which the GP solution must survive) are dynamic, unpredictable and unforgiving. How can GP be improved so that solutions
are produced that are robust to non-trivial changes in the environment? We explore two new approaches. The first approach uses subsets of extreme environments
during training and the second approach uses a voting committee of GP individuals with differing phenotypic behaviour.

  • Content Type Journal Article
  • Pages 1-14
  • DOI 10.1007/s00500-009-0511-4
  • Authors
    • Wei Yan, University College London Department of Computer Science Gower Street London WC1E 6BT UK
    • Christopher D. Clack, University College London Financial Computing, Department of Computer Science Gower Street London WC1E 6BT UK