Search

This paper proposes a novel algorithm for nonnegative independent component analysis, which is based on minimizing the mutual information of the separated signals, and is truly insensitive to the particular underlying distribution of the source data. The unmixing system culminates to a novel neural network model. Compared with other algorithms for nonnegative ICA, the method proposed in this paper can work efficiently even in the case that the source signals are not well grounded, and that pre-whiting process is not needed. Finally, the experiments were performed on both simulating signals and mixtures of image data, the results indicate that the algorithm is efficient and effective. [ABSTRACT FROM AUTHOR]

A Personal Assistant (PA) agent is a software agent capable of helping people to handle tasks in their workplace. The paper proposes a memory mechanism for personal assistant agents in order to enhance agent intelligence while working with the user or with other agents. Inspired by a case memory model in the domain of Case-Based Reasoning (CBR), this paper endows PA agents with a case memory mechanism, which results in improved PA agents: MemoPAs. We present the memory mechanism of MemoPA in detail, and report a first implementation of the method. Finally, future work is outlined for improving the memory mechanism. [ABSTRACT FROM AUTHOR]

A top-down experimental procedure for defect type recognition of epoxy-resin transformers by Partial Discharge (PD) is proposed. Most of the PD detection methods could be performed only at the shutdown period of equipments. By using Acoustic Emission (AE), the real-time and online detection could be reachable. Therefore, this paper conducted high voltage test of pre-faulty transformers and measured those PD signals for recognition needed. Afterward, the selected features that proposed in this paper can be extracted from these collected PD signals. According to these features, effective identification of their faulty types can be done using the proposed particle swarm optimization combined with neural network. Finally, with a view to apply in field, this research adds different noise levels to distort the original data. These distorted data are entered for subsequent testing. Research shows encouraging results that with 30% noise per discharge count, an 80% successful recognition rate can be achieved. [ABSTRACT FROM AUTHOR]

Electromagnetism-like algorithm (EM) is a population-based meta-heuristic which has been proposed to solve continuous problems effectively. In this paper, we present a new meta-heuristic that uses the EM methodology to solve the single machine scheduling problem. Single machine scheduling is a combinatorial optimization problem. Schedule representation for our problem is based on random keys. Because there is little research in solving the combinatorial optimization problem (COP) by EM, the paper attempts to employ the random-key concept enabling EM to solve COP in single machine scheduling problem. We present a hybrid algorithm that combines the EM methodology and genetic operators to obtain the best/optimal schedule for this single machine scheduling problem, which attempts to achieve convergence and diversity effect when they iteratively solve the problem. The objective in our problem is minimization of the sum of earliness and tardiness. This hybrid algorithm was tested on a set of standard test problems available in the literature. The computational results show that this hybrid algorithm performs better than the standard genetic algorithm. [ABSTRACT FROM AUTHOR]

PVM (Parallel virtual machine) library is a tool which used processes large amounts of data sets. This paper wants to achieve a high performance solution that exploits PVM library and parallel computers to solve ICA (Independent Component Analysis) problem. The paper presents parallel power ICA implementations to decomposition data sets. Power iteration (PI) is an algorithm for independent component analysis, which has some desired features. It has higher performance and data capacity than current sequential implementations. This paper, we show the power iteration algorithm which learning updating is in the form of matrix transformation . From power iteration algorithm, we develop parallel power iteration algorithm and implement parallel component decomposition solution. At last, experimental results, analysis and future plans are presented. [ABSTRACT FROM AUTHOR]

Upon close examination of a set of industrial data from a large scale power plant, time varying behavior are discovered. If a fixed model is applied to monitor this process, false alarms will be inevitable. This paper suggests the use of adaptive models to cope with such situation. A recently proposed technique, fast algorithm for Moving Window Principal Component Analysis (MWPCA) was employed because of its following strength: (i) the ability in adapting process changes, (ii) the conceptual simplicity, and (iii) its computational efficiency. Its advances in fault detection is demonstrated in the paper by comparing with the conventional PCA. In addition, this paper proposed to plot the scaled variables in conjunction with MWPCA for fault diagnosis, which is proved to be effective in this application. [ABSTRACT FROM AUTHOR]

In this paper, global exponential stability of Cohen-Grossberg neural networks with reaction-diffusion and Dirichlet boundary conditions is considered by using an approach based on the delay differential inequality and the fixed-point theorem. Some sufficient conditions are obtained to guarantee that the reaction-diffusion Cohen-Grossberg neural networks are globally exponentially stable. The results presented in this paper are the improvement and extension of the existed ones in some existing works. [ABSTRACT FROM AUTHOR]

This paper analyzes global convergence and learning parameters of the back-propagation algorithm for quadratic functions. Some global convergence conditions of the steepest descent algorithm are obtained by directly analyzing the exact momentum equations for quadratic cost functions. In addition, in order to guarantee the convergence for a given learning task, the method is obtained to choose the proper learning parameters. The results presented in this paper are the improvement and extension of the existed ones in some existing works. [ABSTRACT FROM AUTHOR]

Bearing fault detection with the aid of the vibration signals is presented. In this paper, time-domain features are extracted to indicate bearing fault, which collected from tri-axial vibration signal. Decision tree is chosen as an effective diagnostic tool to obtain bearing status. The paper also introduces principal component analysis (PCA) algorithm to reduce training data dimension and remove irrelevant data. Both original data and PCA-based data are used to train C4.5 decision tree models. Then, the result of PCA-based decision tree is compared with normal decision tree to get the best performance of classification process. [ABSTRACT FROM AUTHOR]

This paper proposed a novel residual capacity estimation method for lead-acid battery. Generally, the battery residual capacity is related to open-circuit-voltage (OCV) and inner resistance. Due to complexities, delay, coupling and nonlinearity on the relation described above, the residual capacity estimation is difficult to be obtained by the voltage or inner resistance measurement only. In this paper, by observing and recording the constant current discharging process, we generated the residual capacity pattern for different capacity levels and built a CMAC (Cerebellar Model Articulation Controller) neural network to estimate the lead-acid battery residual capacity. The characteristic of self-learning and generalization, like the cerebellum of human being, a CMAC NN estimation scheme enables powerful, straightforward, and efficient battery residual capacity estimation. With application of this scheme to the experimental data test, the estimation results demonstrate the new scheme with high accuracy and high noise rejection ability. [ABSTRACT FROM AUTHOR]

Driven by new regulations and animal welfare, the need to develop in silico models has increased recently as alternative approaches to safety assessment of chemicals without animal testing. This paper describes a novel machine learning ensemble approach to building an in silico model for the prediction of the Ames test mutagenicity, one of a battery of the most commonly used experimental in vitro and in vivo genotoxicity tests for safety evaluation of chemicals. Evolutionary random neural ensemble with negative correlation learning (ERNE) [1] was developed based on neural networks and evolutionary algorithms. ERNE combines the method of bootstrap sampling on training data with the method of random subspace feature selection to ensure diversity in creating individuals within an initial ensemble. Furthermore, while evolving individuals within the ensemble, it makes use of the negative correlation learning, enabling individual NNs to be trained as accurate as possible while still manage to maintain them as diverse as possible. Therefore, the resulting individuals in the final ensemble are capable of cooperating collectively to achieve better generalization of prediction. The empirical experiment suggest that ERNE is an effective ensemble approach for predicting the Ames test mutagenicity of chemicals. [ABSTRACT FROM AUTHOR]

This paper presents a novel fault detection and section estimation method for unbalanced underground distribution systems (UDS). The method proposed is based on artificial neural networks (ANNs) and wavelet transforms (WTs). The majority of UDS are characterized by having several single/double phase laterals and non-symmetrical lines. Also, Digital Fourier Transforms (DFT), used in the majority of traditional protection relays, supplies a low level of robustness to the fault diagnosis process due to its inversely proportional time-frequency characteristic. These characteristics compromise the traditional fault diagnosis methods performance. ANNs are capable of learning and generalizing, whereas WTs are robust tools capable of evaluating a signal's frequency range that can characterize the fault phenomenon. This paper describes the proposed diagnosis method and discusses the results obtained from simulated implementation. The obtained results demonstrate the capability and robustness of the technique indicating its potential for on-line applications. [ABSTRACT FROM AUTHOR]

This paper presents a Genetic algorithm (GA) based solution to co-optimize test scheduling and wrapper design for core based SOCs. Core testing solutions are generated as a set of wrapper configurations, represented as rectangles with width equal to the number of TAM (Test Access Mechanism) channels and height equal to the corresponding testing time. A locally optimal best-fit heuristic based bin packing algorithm has been used to determine placement of rectangles minimizing the overall test times, whereas, GA has been utilized to generate the sequence of rectangles to be considered for placement. Experimental result on ITC'02 benchmark SOCs shows that the proposed method provides better solutions compared to the recent works reported in the literature. [ABSTRACT FROM AUTHOR]

Kernel methods like support vector machine, kernel principal component analysis and kernel fisher discriminant analysis have recently been successfully applied to solve pattern recognition problems such as face recognition. However, most of the papers present the results without giving kernel parameters, or giving parameters without any explains. In this paper, we present an experiments based approach to optimize the performance of a Gabor feature and kernel method based face recognition system. During the process of parameter tuning, the robustness of the system against variations of kernel function, kernel parameters and Gabor features are extensively tested. The results suggest that the kernel method based approach, with tuned parameters, achieves significantly better results than other algorithms available in literature. [ABSTRACT FROM AUTHOR]

Gyro survey technique has applied and played an important role in many areas, such as offshore oil drilling, directional drilling and so on. Considering the influence of the compensation for the large surveying azimuth error, in this paper, the principle of the gyro survey system is described, and the ANFIS architecture is employed to model the survey azimuth error, predict a chaotic current, all yielding remarkable results according to the gyro survey principle and the data sampled from the two-axis turntable. From the simulation and the testing result, we can see that ANFIS is an effective and feasible way to model and compensate the azimuth error, and the precision of the ANFIS method is higher than the methods of the bilinear interpolation and the radius basis function (RBF), so it is available and advisable in engineering. [ABSTRACT FROM AUTHOR]

Understanding and adapting to changes of customer behavior is an important aspect of surviving in a continuously changing market environment for a modern company. The concept of customer change model mining is introduced and its process is analyzed in this paper. A customer change model mining method based on swarm intelligence is presented, and the strategies of pheromone updating and items searching are given. Finally, an examination on two customer datasets of a telecom company illuminates that this method can achieve customer change model efficiently. [ABSTRACT FROM AUTHOR]

In this paper, a new framework for data analysis based on the "key points" in data distribution is proposed. Here, the key points contain three types of data points: bridge points, border points, and skeleton points, where our main contribution is the bridge points. For each type of key points, we have developed the corresponding detection algorithm and tested its effectiveness with several synthetic data sets. Meanwhile, we further developed a new hierarchical clustering algorithm SPHC (Skeleton Point based Hierarchical Clustering) to demonstrate the possible applications of the key points acquired. Based on some real-world data sets, we experimentally show that SPHC performs better compared with several classical clustering algorithms including Complete-Link Hierarchical Clustering, Single-Link Hierarchical Clustering, KMeans, Ncut, and DBSCAN. [ABSTRACT FROM AUTHOR]

In this paper, we develop a novel approach to perform kernel parameter selection for Kernel Fisher discriminant analysis (KFDA) based on the viewpoint that optimal kernel parameter is associated with the maximum linear separability of samples in the feature space. This makes our approach for selecting kernel parameter of KFDA completely comply with the essence of KFDA. Indeed, this paper is the first paper to determine the kernel parameter of KFDA using a search algorithm. Our approach proposed in this paper firstly constructs an objective function whose minimum is exactly equivalent to the maximum of linear separability. Then the approach exploits a minimum search algorithm to determine the optimal kernel parameter of KFDA. The convergence properties of the search algorithm allow our approach to work well. The algorithm is also simple and not computationally complex. Experimental results illustrate the effectiveness of our approach. [ABSTRACT FROM AUTHOR]

The problem of mean square exponential stability of uncertain stochastic Hopfield neural networks with interval time-varying delays is investigated in this paper. The delay factor is assumed to be time-varying and belongs to a given interval, which means that the derivative of the delay function can exceed one. The uncertainties considered in this paper are norm-bounded and possibly time-varying. By Lyapunov-Krasovskii functional approach and stochastic analysis approach, a new delay-dependent stability criteria for the exponential stability of stochastic Hopfield neural networks is derived in terms of linear matrix inequalities(LMIs). A simulation example is given to demonstrate the effectiveness of the developed techniques. [ABSTRACT FROM AUTHOR]

This paper focuses on two things. One is to describe implementation of a Low-Rate Ultra-Wideband (LR-UWB) transceiver for low power Wireless Personal Area Network (WPAN) applications. The other is to derive the Bit Error Rate (BER) performance of the proposed system. At first, in order to implement low-power and low-complexity transceiver, we design a non-coherent On-Off-Keying(OOK) UWB system by adopting the architecture of the simplified asynchronous transmission and the edge-triggered pulse transmission, which makes the system performance independent of the shape of the transmitted waveform, robust to multipath channels. The designed non-coherent UWB transceiver architecture has an advantage of the simple realization since any mixer, high-speed correlator, and very high-sampling A/D converter are not necessary at the cost of performance degradation of about 3dB. Further, the designed non-coherent UWB transceiver is actually implemented with the wireless CANVAS prototyping testbed in short range indoor application environment such as a lecture room. The implemented prototype testbed is proven to offer the data rate of 9.6kbps and 115kbps on the conditions of Peer-to-Peer (P-to-P) in the indoor channel within the range of about 10m. At second, we analyze the bit error rate (BER) performance of this proposed system under the line-of-sight (LOS) residential channel environment. As a result, it is necessary to choose appropriate number of Window Bank(WB)s for better BER performance. [ABSTRACT FROM AUTHOR]

Hybrid actuators, which are a combination of two types of motor and mechanism, have good flexibility. In this paper, the kinematics analysis for a hybrid five-bar actuator is introduced. An optimization design of hybrid actuator is performed with reference to kinematics objective function. By the use of the properties of ergodicity, stochastic property, and "regularity" of chaos, a hybrid optimization algorithm is proposed based on chaos optimization algorithm (COA) and a gradient-based search. The efficiency of COA is much higher than some stochastic algorithms such as simulated anneal algorithm (SAA) and genetic algorithm (GA) when COA is used to a kind of continuous problems. The chaos optimization algorithm can improve the efficiency of searching in the whole field by gradually shrinking the area of optimization variable. The precision of optimum dimensions obtained by using the hybrid optimization method can be improved evidently. Finally, a numerical example is carried out, and the simulation results show that the optimization method is feasible and satisfactory in the design of hybrid actuator. [ABSTRACT FROM AUTHOR]

In this paper, a difference scheme satisfying two conservative laws for the periodic Camassa-Holm equation is given. Convergence is proved. The numerical simulation exhibits the time evolution and the interaction of solitary wave solutions. [ABSTRACT FROM AUTHOR]

Selectivity estimation is one of the query optimization techniques. It is difficult for the previous selectivity estimation techniques for moving objects to apply the location change of moving objects to synopsis. Therefore, they result in much error when estimating selectivity for queries, because they are based on the extended spatial synopsis which does not consider the property of the moving objects. In order to reduce the estimation error, the existing techniques should often rebuild the synopsis. Consequently problem occurs, that is, the whole database should be read frequently. In this paper, we proposed a moving object histogram method based on quadtree to develop a selectivity estimation technique for moving object queries. We then analyzed the performance of the proposed method through the implementation and evaluation of the proposed method. Our method can be used in various location management systems such as vehicle location tracking systems, location based services, telematics services, emergency rescue service, etc in which the location information of moving objects changes over time. [ABSTRACT FROM AUTHOR]

Conditional density estimation is very important in financial engineer, risk management, and other engineering computing problem. However, most regression models have a latent assumption that the probability density is a Gaussian distribution, which is not necessarily true in many real life applications. In this paper, we give a framework to estimate or predict the conditional density mixture dynamically. Through combining the Input-Output HMM with SVM regression together and building a SVM model in each state of the HMM, we can estimate a conditional density mixture instead of a single gaussian. With each SVM in each node, this model can be applied for not only regression but classifications as well. We applied this model to denoise the ECG data. The proposed method has the potential to apply to other time series such as stock market return predictions. [ABSTRACT FROM AUTHOR]

In this paper, we propose a feature extraction method and fusion algorithm which is constructed by PCA and LDA to detect a fault state of the induction motor that is applied over the whole field of a industry. After yielding a feature vector from current signal which is measured by an experiment using PCA and LDA, we use the reference data to produce matching values. In a diagnostic step, two matching values which are respectively obtained by PCA and LDA are combined by probability model, and a faulted signal is finally diagnosed. As the proposed diagnosis algorithm brings only merits of PCA and LDA into relief, it shows excellent performance under the noisy environment. The simulation is executed under various noisy conditions in order to demonstrate the suitability of the proposed algorithm and it showed more excellent performance than the case just using conventional PCA or LDA [ABSTRACT FROM AUTHOR]

As a practical generalization of the job shop scheduling problem, multi-resource job shop scheduling problem (MRJSSP) is discussed in this paper. In this problem, operations may be processed by a type of resources and jobs have individual deadlines. How to design and optimize this problem with DSAFO, a novel multi-agent algorithm, is introduced in detail by a case study, including problem analysis, agent role specification, and parameter selection. Experimental results show the effectiveness and efficiency of designing and optimizing MRJSSPs with multi-agent. [ABSTRACT FROM AUTHOR]

Multivariate visualization techniques are often used as assistant tools for classification tasks up to now. However, few classification systems fully utilize the capability of multivariate visualization and integrate them with multivariate analysis algorithms into a compact system. We propose an interactive visual classification model based on some multivariate graphical presentation in this paper. As an example of it, a visual classifier based on parallel coordinates plot is developed. The multivariate data is first mapped to the parallel coordinates plot, and then an optimizer based on linear discriminant analysis optimizes it into the visualization more fit for classification tasks. This optimized visualization then can be processed by decision tree algorithm and attain classification rules. It has the merit of making the invisible visible and users can steer the classification process, consequently favor the understanding and knowledge discovery of original data. [ABSTRACT FROM AUTHOR]

The goal of image registration is to align two or more images of the same scene. To fully automate the registration process by optimization of the mutual information criterion, a robust global optimizer is indispensable. This paper focuses mainly on application and evaluation of particle swarm optimization for the rigid transformation registration of multimodality images. Four different modes of particle swarm optimization are proposed to globally optimize the challenging intensity based image registration. The functional manifest of different modes is comparatively evaluated by rigid experiments. The results show that the particle swarm optimization algorithm is to be promising for optimal multimodality image registration. With consideration of the trade-off between the successful rate and the computation time, the proposed PSO Mode II gives relatively best performance during the experimental studies of multimodality image registration. [ABSTRACT FROM AUTHOR]

This paper presents a new autonomous kinematic calibration technique by using a laser-vision sensor called "Perceptron TriCam Contour". Because the sensor measures by capturing the image of a projected laser line on the surface of the object, we set up a long, straight line of a very fine string inside the robot workspace, and then allow the sensor mounted on a robot to measure the point intersection of the line of string and the projected laser line. The data collected by changing robot configuration and measuring the intersection points are constrained to on a single straght line such that the closed-loop calibration method can be applied. The obtained calibration method is simple and accurate and also suitable for on-site calibration in an industrial environment. The method is implemented using Hyundai VORG-35 for its effectiveness. [ABSTRACT FROM AUTHOR]

A text prompted speaker verification system is presented in this paper. This system is based on ten Chinese digits. Basic acoustic models are speaker dependent and content dependent phoneme HMMs which were generated by adapting speaker independent models to the utterances of specific speakers. An obvious constraint for normalization techniques used in TDSV is that the phrases with the same content should be used for competitive cohort models. So many of the score normalization techniques are either difficult to implement because of lack of data or not good for performance improvement because of poor estimation of the normalization parameters. We propose a method which combines the traditional T-Norm and Cohort Norm together to find a good tradeoff of testing utterance normalization and target speaker model normalization. The proposed method improved the system performance from the baseline equal error rate 3.42% for T-Norm and 2.72% for Cohort Norm to 2.50%. [ABSTRACT FROM AUTHOR]

In this paper, we developed a biometric user authentication system based on 3-dimensional (3D) face recognition under ubiquitous computing environment. Since 2D based face recognition has been shown its structural limitation, 3D model based approach for face recognition has been spotlighted as a robust solution under variant conditions of pose and illumination. Since 3D face model consists of a large number of vertices, 3D model based face recognition system is generally inefficient for real-time computation. We propose a novel 3D face representation algorithm to reduce the number of vertices and optimize its computation time while maintaining reasonable recognition performance. We evaluate the performance of proposed algorithm with the Korean face database collected using a stereo-vision based 3D face capturing device. Additionally, various decision making similarity measures were explored for recognition performance. Our experimental results indicated that our proposed algorithm is robust for biometric user authentication and is also reasonably fast for real-time processing. [ABSTRACT FROM AUTHOR]

A trend actually characterizing any technological and application field is the use of control and computing devices. This moves the attention of the insiders on the management of more complex processes and systems, modifying requirements, tasks and skills. In this scenario the dependability evaluation becomes strategic. The increasing complexity of systems requires a serious revision of the techniques and the methodologies exploited for evaluating the system dependability. Notations as reliability block diagrams (RBD), fault trees (FT) or reliability graphs (RG) and their extensions, such as dynamic FT (DFT), could not adequately model dynamic reliability/availability aspects and behaviors as the units' interactions. To overcome those problems we developed a new formalism derived from RBD: the dynamic RBD (DRBD). DRBD allow to represent dynamic reliability/availability behaviors by specifying the dependency concept. The dependency is the building block for modeling dynamic aspects in DRBD. In this paper we detail the DRBD modeling technique thorough an example of distributed computing system taken from literature. By this we also compare our approach with the DFT one, providing proofs of the DRBD potentialities in dynamic reliability/availability modeling. [ABSTRACT FROM AUTHOR]

For finding the actual diagnosis of the faulty system, this paper discusses the relationship between the candidate of diagnosis and the set of the actual faulty components. Then we define the notion of adoptability of the diagnostic system and prove that the consistency-based diagnosis is adoptable. On the basis of the above works, a test theory of the diagnosis based on consistency is proposed, which indicates that how tests provide information about the current space of diagnoses. [ABSTRACT FROM AUTHOR]

KGFST (Kernel Generalized Foley-Sammon Transform) has been proved very successfully in the area of pattern recognition. By the kernel trick, one can calculate KGFST in input space instead of feature space to avoid high dimensional problems. But one has to face two problems. In many applications, when n (the number of samples) is very large, it not realistic to store and calculate serval n×n metrics. Another problem is the complexity for the eigenvalue problem of n×n metrics is O(n3). So a new nonlinear feature extraction method CW-KGFST (KGFST with Cluster-weighted) based on KGFST and Clustering is proposed in this paper. Through Cluster-weighted, the number of samples can be reduced, the calculate speed can be higher and the accuracy can be preserved simultaneously. Lastly, our method is applied to digits and images recognition problems, and the experimental results show that the performance of present method is superior to the original method. [ABSTRACT FROM AUTHOR]

Support vector machines (SVMs) were designed for two-class classification problems, and multi-class classification problems have been solved by combining independently produced two-class decision functions. In this paper, we propose two multi-class Lagrangian Support Vector Machine(LSVM) algorithms using the quick and simple properties of LSVM. The experimental results in the linear and nonlinear cases indicate that the CPU running time of these two algorithms is shorter than that of the standard support vector machines, and their training correctness and testing correctness are almost identical. [ABSTRACT FROM AUTHOR]

The fermentation process is very complex and non-linear, many parameters are not easy to measure directly on line, soft sensor modeling is a good solution. This paper introduces v-support vector regression (v-SVR) for soft sensor modeling of fed-batch fermentation process. v-SVR is a novel type of learning machine. It can control the accuracy of fitness and prediction error by adjusting the parameter v. An on-line training algorithm is discussed in detail to reduce the training complexity of v-SVR. The experimental results show that v-SVR has low error rate and better generalization with appropriate v. [ABSTRACT FROM AUTHOR]

Nearest Neighbor (NN) rule is one of the simplest and most important methods in pattern recognition. In this paper, we propose a kernel difference-weighted k-nearest neighbor method (KDF-WKNN) for pattern classification. The proposed method defines the weighted KNN rule as a constrained optimization problem, and then we propose an efficient solution to compute the weights of different nearest neighbors. Unlike distance-weighted KNN which assigns different weights to the nearest neighbors according to the distance to the unclassified sample, KDF-WKNN weights the nearest neighbors by using both the norm and correlation of the differences between the unclassified sample and its nearest neighbors. Our experimental results indicate that KDF-WKNN is better than the original KNN and distance-weighted KNN, and is comparable to some state-of-the-art methods in terms of classification accuracy. [ABSTRACT FROM AUTHOR]

Blood oxygenation level dependent (BOLD)contrast based functional magnetic resonance imaging (fMRI)can be used to detect brain neural activities. In this paper, a new procedure is presented which allows the estimation of the hemodynamic approach from BOLD responses. The procedure is based on Friston proposed dynamic model and Agnes Aubert proposed a correlation model between activation and motabolism, in this case, adopted to characterize hemodynamic responses in functional magnetic resonance imaging (fMRI). This work represents a fundamental improvement over existing approaches to system identification using nonlinear hemodynamic models. The model can simulate the change of oxygen motabolism, de-oxyhemoglobine, cerebral blood flow and volume to brain activation. [ABSTRACT FROM AUTHOR]

Through the situation today it is necessary that small and medium sized enterprises collaborate in so-called collaborative operation network. In the center of interest is the development of a virtual enterprise model which is based on small collaborative cells in so-called operation centers. Thus, the concentration on the core competences is supported and the market power is increased by the help of the collaborative operation networks. The automated selection of the partners is the major problems in virtual enterprises. In the paper, a method for choosing the most capable operation centers for every order is designed. The selected operation centers fulfill the tasks of a value chain particularly well. Within the approach, the problem will be solved by Ant Colony Optimization in combination with the Analytical Hierarchy Process. [ABSTRACT FROM AUTHOR]

In this paper, a new methodology for medical diagnosis based on fuzzy clustering and parallel neural networks is proposed. Intelligent systems have various fields. Breast cancer is one of field to be targeted, which is the most common tumor-related disease among women. Diagnosis of breast cancer is not task for medical expert owing to many attributes of the disease. So we proposed a new method, FCM based parallel neural networks to handle difficult. FCM based parallel neural networks composed of two parts. One is classifying breast cancer data using Fuzzy c-means clustering method (FCM). The other is designing the multiple neural networks using classified data by FCM. The proposed methodology is experimented, evaluated, and compared the performance with other existed models. As a result we can show the effectiveness and precision of the proposed method are better than other previous models. [ABSTRACT FROM AUTHOR]

Hardware/Software (HW/SW) partitioning is becoming an increasingly crucial step in embedded system co-design. To cope with roughly assumed parameters in system specification and imprecise benchmarks for judging the solution's quality, researchers have been trying to find methods for a semi-optimal partitioning scheme in HW/SW partitioning for years. This paper proposes an application of a hybrid neural fuzzy system incorporating Boltzmann machine to the HW/SW partitioning problem. The hybrid neural fuzzy system's architecture and performance estimation against simulated annealing algorithm are evaluated. The simulation result shows the proposed system outperforms other algorithm in cost and performance. [ABSTRACT FROM AUTHOR]

The buck converter is one of DC/DC converters that are often used as power supplies. This paper presents parameter tuning methods to obtain circuit element values for the buck converter to minimize the output voltage variation under load changing environments. The conventional method using the concept of the phase margin is extended to have optimal phase margin that gives slightly improved performance in the output voltage response. For this, the phase margin becomes the tuning parameter that is optimized with the genetic algorithm. Next, the circuit element values are directly considered as the tuning parameters and optimized using the genetic algorithm to have very improved performance in the output voltage control of the buck converter. [ABSTRACT FROM AUTHOR]

As a novel evolutionary computing technique, Differential Evolution (DE) has been considered to be an effective optimization method for complex optimization problems, and achieved many successful applications in engineering. In this paper, a new algorithm of Quadratic Partial Least Squares (QPLS) based on Nonlinear Programming (NLP) is presented. And DE is used to solve the NLP so as to calculate the optimal input weights and the parameters of inner relationship. The simulation results based on the soft measurement of diesel oil solidifying point on a real crude distillation unit demonstrate that the superiority of the proposed algorithm to linear PLS and QPLS which is based on Sequential Quadratic Programming (SQP) in terms of fitting accuracy and computational costs. [ABSTRACT FROM AUTHOR]

This paper introduces the main ideas of the nested partitions (NP) method, analyses its efficiency theoretically and proposes the way to improve the optimization efficiency of the algorithm. Then the paper introduces the simulated annealing (SA) algorithm and incorporates the ideas of SA into two of the arithmetic operators of NP algorithm to form the combined NP/SA algorithm. Moreover, the paper presents the explicit optimization procedure of the combined algorithm NP/SA and explains the feasibility and superiority of it. The NP/SA algorithm adopts the global optimization ability of NP algorithm and the local search ability of SA algorithm so that it improves the optimization efficiency and the convergence rate. This paper also illustrates the NP/SA algorithm through an optimization example. [ABSTRACT FROM AUTHOR]

Image segmentation is a fundamental step in image processing. Otsu's threshold method is a widely used method for image segmentation. In this paper, a novel image segmentation method based on chaos immune clone selection algorithm (CICSA) and Otus's threshold method is presented. By introducing the chaos optimization algorithm into the parallel and distributed search mechanism of immune clone selection algorithm, CICSA takes advantage of global and local search ability. The experimental results demonstrate that the performance of CICSA on application of image segmentation has the characteristic of stability and efficiency. [ABSTRACT FROM AUTHOR]

Clustering is one of the most important tasks in data mining. The K-means algorithm is the most popular one for achieving this task because of its efficiency. However, it works only on numeric values although data sets in data mining often contain categorical values. Responding to this fact, the K-modes algorithm is presented to extend the K-means algorithm to categorical domains. Unfortunately, it suffers from computing the dissimilarity between each pair of objects and the mode of each cluster. Aiming at addressing these problems confronting K-modes, we present a new algorithm called K-distributions in this paper. We experimentally tested K-distributions using the well known 36 UCI data sets selected by Weka, and compared it to K-modes. The experimental results show that K-distributions significantly outperforms K-modes in term of clustering accuracy and log likelihood. [ABSTRACT FROM AUTHOR]

Collaborative Filtering(CF) is a widely accepted method of creating recommender systems. CF is based on the similarities among users or items. Measures of similarity including the Pearson Correlation Coefficient and the Cosine Similarity work quite well for explicit ratings, but do not capture real similarity from the ratings derived from implicit feedback. This paper identifies some problems that existing similarity measures have with implicit ratings by analyzing the characteristics of implicit feedback, and proposes a new similarity measure called Inner Product that is more appropriate for implicit ratings. We conducted experiments on user-based collaborative filtering using the proposed similarity measure for two e-commerce environments. Empirical results show that our similarity measure better captures similarities for implicit ratings and leads to more accurate recommendations. Our inner product-based similarity measure could be useful for CF-based recommender systems using implicit ratings in which negative ratings are difficult to be incorporated. [ABSTRACT FROM AUTHOR]

To resolve the shortage of traditional clustering algorithm when dealing data set with complex distribution, a novel adaptive k-Nearest Neighbors clustering(AKNNC) algorithm is presented in this paper. This algorithm is made up of three parts: (a)normalize data set; (b)construct initial patterns; (c)merge initial patterns. Simulation results show that compared with classical FCA, our AKNNC algorithm not only has better clustering performance for data set with Complex distribution, but also can be applied to the data set without knowing cluster number in advance. [ABSTRACT FROM AUTHOR]

This paper presents a novel Wiener-type recurrent neural network with the minimum description length (MDL) principle for unknown dynamic nonlinear system identification. The proposed Wiener-type recurrent network resembles the conventional Wiener model that consists of a dynamic linear subsystem cascaded with a static nonlinear subsystem. The novelties of our approach include: 1) the realization of a conventional Wiener model into a simple connectionist recurrent network whose output can be expressed by a nonlinear transformation of a linear state-space equation; 2) the state-space equation mapped from the network topology can be used to analyze the characteristics of the network using the well-developed theory of linear systems; and 3) the overall network structure can be determined by the MDL principle effectively using only the input-output measurements. Computer simulations and comparisons with some existing recurrent networks have successfully confirmed the effectiveness and superiority of the proposed Wiener-type network with the MDL principle. [ABSTRACT FROM AUTHOR]

As a power tool for analyzing data, concept lattice has been extensively applied in several areas such as knowledge discovery, software engineering and case-based reasoning. However, building concept lattice is time-consuming and complicated; it becomes the bottleneck of application. Therefore, a simple and efficient method for building concept lattice is proposed in this paper. We transform binary formal context into matrix at first, and then discuss how to build concept lattice based on basic concepts and added concepts, which the two concepts can be got from matrix operation. We also present a fast algorithm BCLMO (Building Concept Lattice based on Matrix Operation) for building concept lattice, and analyze the time complexity of BCLMO. The method we proposed could remarkably reduce the time complexity and improve the efficiency of building concept lattice. [ABSTRACT FROM AUTHOR]