By Topic

Computational Intelligence and Industrial Application, 2008. PACIIA '08. Pacific-Asia Workshop on

Date 19-20 Dec. 2008

Go

Filter Results

Displaying Results 1 - 25 of 219
  • [Front cover - Vol 2]

    Page(s): C1
    Save to Project icon | Request Permissions | PDF file iconPDF (1962 KB)  
    Freely Available from IEEE
  • [Title page i - Volume 2]

    Page(s): i
    Save to Project icon | Request Permissions | PDF file iconPDF (28 KB)  
    Freely Available from IEEE
  • [Title page iii - Volume 2]

    Page(s): iii
    Save to Project icon | Request Permissions | PDF file iconPDF (119 KB)  
    Freely Available from IEEE
  • [Copyright notice - Volume 2]

    Page(s): iv
    Save to Project icon | Request Permissions | PDF file iconPDF (148 KB)  
    Freely Available from IEEE
  • Table of contents - Volume 2

    Page(s): v - xix
    Save to Project icon | Request Permissions | PDF file iconPDF (146 KB)  
    Freely Available from IEEE
  • Message from Workshop Chairs - Volume 2

    Page(s): xx
    Save to Project icon | Request Permissions | PDF file iconPDF (166 KB)  
    Freely Available from IEEE
  • PACIIA 2008 Organizing Committee - Volume 2

    Page(s): xxi
    Save to Project icon | Request Permissions | PDF file iconPDF (227 KB)  
    Freely Available from IEEE
  • PACIIA 2008 Committee Members - Volume 2

    Page(s): xxii
    Save to Project icon | Request Permissions | PDF file iconPDF (169 KB)  
    Freely Available from IEEE
  • Distributed and Parallelled EM Algorithm for Distributed Cluster Ensemble

    Page(s): 3 - 8
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (282 KB) |  | HTML iconHTML  

    The paper introduces base clusterings distributed cluster ensemble which can handle the problems of privacy preservation, distributed computing and knowledge reuse. First, the latent variables in latent Dirichlet location model for cluster ensemble (LDA-CE) are defined and some terminologies are defined. Second, Variational approximation inference for LDA-CE is stated in detail. Third, base on the variational approximation inference, we design a distributed and paralleled EM algorithm for cluster ensemble (DPEM). Finally, some datasets from UCI are chosen for experiment, Compared with cluster-based similarity partitioning algorithm (CSPA), hyper-graph partitioning algorithm(HGPA) and meta-clustering algorithm(MCLA), the results show DPEM algorithm does work better and DPEM can work distributed and paralleled, so DPEM can protect privacy information more and can save time. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Study of Index Weight in Network Threat Evaluation Based on Improved Grey Theory

    Page(s): 9 - 13
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (395 KB) |  | HTML iconHTML  

    It's necessary to create a credible and general evaluating index model in network threat assessment. According to the network indices' effect, this paper proposed an algorithm which combines grey correlation degree in grey theory with traditional analytic hierarchy process (AHP) to determine the indices. It considered both objective and subjective factors. Experiments showed that this method had a strict mathematical theory basis, a definite meaning and a high practical value. It improved the evaluation methods' credibility of network threat assessment. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Blog Community Discovery Based on Tag Data Clustering

    Page(s): 14 - 18
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (374 KB) |  | HTML iconHTML  

    Blog is increasingly becoming an important source of information. Blog community is a kind of a group of bloggers with the same interest and common topics on the Internet. To use blog resources effectively, one important way is to identify blog communities and their members in order to refine the blog circle. In this paper, we first define the blog community and the community center, and then construct the blog community discovery algorithm, and next, we collect and analyze the blog tag data from "sina blogsphere". Through the establishment of "blogger- frequency" matrix, we use the clustering method of data mining to implement the discovery of blog communities. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • RBF Neural Networks and Its Application in the Simulation of a Random Vibration System

    Page(s): 19 - 22
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (424 KB) |  | HTML iconHTML  

    Radial basis function neural network (RBF-NN) is applied to simulate a random vibration system in this paper. First, an overview of random vibration system is introduced, and the dynamic model of the system and the spectral density of external excitation are presented. Then, a radial basis function neural network is proposed to predict the random vibration based on a biological neuron system. Radial basis Gaussian function and back-propagation learning algorithm are employed to train the proposed NN. The back-propagation algorithm updates the weights and thresholds of the RBF-NN are deduced in detail. At last, the RBF-NN is trained and used to predict the acceleration of random vibration system in different conditions. The simulation results show the advantages of radial basis Gaussian network in fast convergence of the results of different approaches, and the proposed RBF-NN can be used to simulate a random vibration system. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • AND/OR Tree Search Algorithm in Web Service Composition

    Page(s): 23 - 27
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (338 KB) |  | HTML iconHTML  

    Composition of web services has received much interest to support business-to-business or enterprise application integration. Web service composition lets developers combine single services as a composite one to achieve complex goal. It is a highly complex task and already beyond the human capability to deal with the whole process manually, so AI-based approach for automated web service composition has drawn much research attention. AI-based service composition uses some algorithm in Artificial Intelligence, such as AI planning and graph searching. Based on analysis of the limitation of those two approaches, an AND/OR tree-based approach for web service composition is proposed in this paper which is based on Artificial Intelligence. The Input/Output Dependency AND/OR Tree is represented first, and then AND/OR tree search algorithm including basic search algorithm and special algorithm is discussed. With that, an implementation is provided to described whole process when apply AND/OR Tree search algorithm in web service composition field. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Simulation of Critical Flux through Adiabatic Capillary Tubes Based on Artificial Neural Network

    Page(s): 28 - 30
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (334 KB) |  | HTML iconHTML  

    Evaluating the critical flux in capillary tubes is the key to the research on flow characteristics in capillary tubes. The mathematical model of refrigerant flow through the capillary tubes was presented. The numerical solutions were obtained based on the program made. Data derived from capillary tube theoretical models were used as example collection to train the back propaganda(BP)network model in order to evaluate the critical flux through capillary tubes. The results are satisfactory. Compared with finite difference numerical computation method, the computation of the critical flux based on neural network is more useful to engineering design of capillary tubes. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Software Reliability Multi-Scale Prediction Model Based on EMD and RBF Network

    Page(s): 31 - 35
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (245 KB) |  | HTML iconHTML  

    Aiming at the prediction precision and applicability problem for the traditional software reliability prediction models, from the point of nonlinear time sequence, this paper presented a novel software reliability prediction model using RBF neural network based on empirical mode decomposition theory. In the paper, the fault data series obtained from software reliability test phase is decomposed into a series of intrinsic mode functions and a residue signal. Then a RBF network is constructed for an intrinsic mode function or the residual signal. Finally output of every prediction model is integrated into one output with equal weighted. Experimental results showed that the proposed model had higher precision of prediction and better applicability, compared with traditional software reliability models. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A Dynamic Scheduling Algorithm for Distributed Kahn Process Networks in a Cluster Environment

    Page(s): 36 - 42
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (400 KB) |  | HTML iconHTML  

    In this paper, a novel dynamic task scheduling algorithm is proposed for parallel applications modeled in Kahn process networks (KPN) running in a distributed multi-processor cluster. Static job scheduling algorithms do not work for the purpose for that the complexity of a KPN model remains unpredictable at compile time. Dynamic load balancing strategies ignore the explicit data dependences among tasks and may lead to inappropriate process migrations. The algorithm presented in this paper is based on the sequence of dynamic recorded events of each task at runtime. It then predicts the execution efficiency of a KPN model in various scheduling (task-processor assignments) through the estimation of average resource utilization rate. Simulations have shown satisfying results. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The Estimation of the New Feature Extracting Method for Tamper Detection in Printed Documents

    Page(s): 43 - 47
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (397 KB) |  | HTML iconHTML  

    In this paper, we report the experimental result of the estimation of our proposed technique to extract the feature of the printed documents for tamper detection. We have compared the proposed technique which uses median points as a feature and the existing technique which uses the area of the black dots as a feature. As an experiment, we calculated the median point of alphabets and numbers, and printed and scanned to see the probability of the collision, the uniformity of the distribution, invariability during D/A and A/D transform, and the invariability during the ordinal change of paper of our feature extracting method. For the probability of the collision, the proposed method showed lower probability of the collision than the existing one. For the uniformity of the distribution, the proposed method showed a half skew of the existing one. For the invariability during D/A and A/D transform, the proposed method showed the invariability of about a thirtieth of the one of the existing one. For the invariability during the ordinal change of paper, the proposed method showed an eighteenth invariability of the existing one. From the above results, we have concluded that as a feature extracting method, our proposed method is better than the existing one. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Condition Monitoring of Equipment Using a Joint RSAR and Fuzzy ART Neural Network Method

    Page(s): 48 - 52
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (343 KB) |  | HTML iconHTML  

    Working conditions are monitoring parameters are huge and neural network learning time too long in the condition monitoring of multi word condition equipment. To improve monitoring efficiency, a joint rough set attribute reduction (RSAR) and Fuzzy ART (adaptive resonance theory) neural network method is proposed in this study. The dimension of an input vector to Fuzzy ART neural networks can be reduced through RSAR. The updated vectors are used to train Fuzzy ART neural networks. An example is investigated to evaluate the proposed method in this study. Analysis results indicate that the proposed method can save great learning time without losing monitoring capability. Additionally, sensor abnormality and signal transmission issues may be detected as well. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Web Document Query Optimization Based on Memetic Algorithm

    Page(s): 53 - 56
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (326 KB) |  | HTML iconHTML  

    To efficiently retrieve relevant documents from the explosive growth of the Internet and other sources of information access, an efficient document query optimization approach based on memetic algorithm (MA) is proposed. Experimental results show that the proposed algorithm can improve the precision of document retrieval compared with other conventional query optimization algorithm. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A Method of Estimating Network Reliability Using an Artificial Neural Network

    Page(s): 57 - 60
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (362 KB) |  | HTML iconHTML  

    This paper presents a method to estimate the all-terminal reliability of network by neural networks. We first employ the scheme that a network topology is mapped into a binary vector, and use Monte Carlo simulation to obtain sample data of network reliability. Then the neural networks are constructed, trained and validated with the network topologies, links reliabilities and data set of network reliability. A grouped cross-validation approach is adopted to improve the performance of neural networks. The results show the model can carry out the non-linear mapping relationship between the topological structure, the links reliabilities and the reliability of network. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Research on Case Retrieval Method in the Conceptual Design Based on Rough Set

    Page(s): 61 - 65
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (408 KB) |  | HTML iconHTML  

    A rough set-based case retrieval method was presented in intelligent CAD system of the conceptual design. The case knowledge representation model of the conceptual design based on rough set was brought forward by analyzing features of CAD systems. An approach of reducing the feature attributes and calculating the feature attributes weighting was proposed based on rough set theory. The continuous feature attributes were discretized by using hybrid clustering method. Then the importance of the feature attributes were calculated first, and then the importance were normalized to weighting coefficients. It obtains the closest conceptual design case from the case database to take the design case as reference. The proposed method is also demonstrated by an application example. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A Novel Clustering Algorithm with Ant Colony Optimization

    Page(s): 66 - 69
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (499 KB) |  | HTML iconHTML  

    Clustering analysis is an important area of data mining. A kind of new clustering algorithm with ant colony optimization based on cluster center initialization is proposed in this paper. The new algorithm gives initialized cluster centers by different methods, then solves clustering problems by iterated method. Three methods of cluster center initialization are used in clustering algorithm with ant colony optimization - Sacc.Three datasets - butterfly data, iris and wine are chosen for the compare of three algorithms. The results of several times experiments show that the new algorithm is less in running time, is better in clustering effect and more stable than Sacc. Experimental results validate new algorithm's efficiency. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Construction Method of Knowledge Base Based on Fuzzy and Modular Ontology

    Page(s): 70 - 74
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (581 KB) |  | HTML iconHTML  

    Ontology construction is a time consuming and labor intensive task and it is difficult to reuse existed ontology, so the conventional knowledge base based ontology building methods are in a low efficiency. From the view of facilitating knowledge sharing and reusing, modularization is introduced into ontology. But most current methods of ontology modularization are restricted to man-made factors to a large extent. Thus, a novel approach for semi-automatic decomposition of ontology into modules, which takes fuzzily computing concept semantic similarity based ontology as the base, is proposed in this paper. Meanwhile, knowledge base based fuzzy and modular ontology is established. Finally, through a experiment on Grab Ontology the method's validity is proved. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Contextual Hidden Markov Tree Model for the Dual-Tree Complex Wavelet Transform

    Page(s): 75 - 79
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (577 KB) |  | HTML iconHTML  

    Multiresolution models such as the wavelet domain hidden Markov tree (HMT) provide a powerful approach for image modeling and processing because of the persistence properties of wavelet coefficients. In this paper, a new HMT model based on the dual-tree complex wavelet transform is proposed. The model is extended from the contextual hidden Markov tree (CHMT) to the complex wavelet transform, which can mitigate the two problems (shift-variance and lack of the clustering property of wavelet coefficients within a scale) of the conventional wavelet domain HMT model simultaneously. We demonstrate the effectiveness of the model for image denoising. Experiments show that this new model achieves better performance than other related HMT-based image denoising algorithms. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Research on Long Term Load Forecasting Based on Improved Genetic Neural Network

    Page(s): 80 - 84
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (372 KB) |  | HTML iconHTML  

    Considering the features of long term load forecasting are complicated, this paper proposes a generic neural network model that is able to adapt to and learn from amount of non-linear or imprecise rules, so it is a model with highly robustness. For avoiding the inflexibility of the generic neural network itself, many experiences and opinions of experts are introduced during the use, so that a comprehensive effect of different factors that influence the power load can be reflected. The generic algorithm is able to search precisely at global scope, and the neural network is able to fit well at local scope, both of which are chosen together by this paper, i.e. the generic neural network algorithm. In the simulation training of the model, data from 1990 to 2007 of 16 indexes are used, and experience and opinions of national regulation factors of primary, secondary and tertiary industries that are provided by experts are also followed, and the long term load forecasting is carried out by rolling annual extrapolation. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.