• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.28 no.7C
• /
• pp.705-711
• /
• 2003

This paper studies the application of a fuzzy-ARTMAP neural network to digital communications channel equalization. This approach provides new solutions for solving the problems, such as complexity and long training, which found when implementing the previously developed neural-basis equalizers. The proposed fuzzy-ARTMAP equalizer is fast and easy to train and includes capabilities not found in other neural network approaches; a small number of parameters, no requirements for the choice of initial weights, automatic increase of hidden units, no risk of getting trapped in local minima, and the capability of adding new data without retraining previously trained data. In simulation studies, binary signals were generated at random in a linear channel with Gaussian noise. The performance of the proposed equalizer is compared with other neural net basis equalizers, specifically MLP and RBF equalizers.

• Journal of the Korean Data and Information Science Society
• /
• v.11 no.2
• /
• pp.269-277
• /
• 2000

We propose the Bayesian testing for the equality of two log-normal population means. Specifically we use the intrinsic Bayes factors suggested by Berger and Perichi (1996, 1998) based on the noninformative priors for the parameters. In order to investigate the usefulness of the proposed Bayesian testing procedures, we compare it with classical tests via both real data analysis and simulation.

• Proceedings of the KIEE Conference
• /
• 2002.07d
• /
• pp.2034-2036
• /
• 2002

In this paper, we propose the modeling of a chaotic nonlinear system using wavelet neural networks. In our modeling, we used the parameter adjusting method as the training method of a wavelet neural network. The difference between the actual output of a nonlinear chaotic system and that of a wavelet neural network adjusts the parameters of a wavelet neural network using the gradient-descent method. To verify the efficiency of this paper, we perform the simulation using Duffing system, which is a representative continuous time chaotic nonlinear system.

• International Journal of Control, Automation, and Systems
• /
• v.6 no.4
• /
• pp.559-570
• /
• 2008

This paper proposes an improved particle swarm optimization named PSO with Controllable Random Exploration Velocity (PSO-CREV) behaving an additional exploration behavior. Different from other improvements on PSO, the updating principle of PSO-CREV is constructed in terms of stochastic approximation diagram. Hence a stochastic velocity independent on cognitive and social components of PSO can be added to the updating principle, so that particles have strong exploration ability than those of conventional PSO. The conditions and main behaviors of PSO-CREV are described. Two properties in terms of "divergence before convergence" and "controllable exploration behavior" are presented, which promote the performance of PSO-CREV. An experimental method based on a complex test function is proposed by which the proper parameters of PSO-CREV used in practice are figured out, which guarantees the high exploration ability, as well as the convergence rate is concerned. The benchmarks and applications on FCRNN training verify the improvements brought by PSO-CREV.

• Proceedings of the Computational Structural Engineering Institute Conference
• /
• 1995.10a
• /
• pp.95-102
• /
• 1995

Procedures for estimation of axial or flexural rigidities of bridge members by neural networks are shown. To treat large scale structures containing many unkwon parameters, substructuring concept is introduced. The measurement points are selected considering the sensitivity of the element stiffnesses of interest. Utilization of relative mode vectors is found to be very effective for the local parameter estimation. Then, the study focuses on the method to obtain the training set enough to represent structures. It is shown that noise injection is effective to reduce the estimation errors caused by measurement noise. Verification of the present method is carried out using a cable-stayed bridge model.

• Journal of Integrative Natural Science
• /
• v.4 no.4
• /
• pp.266-272
• /
• 2011

Focal adhesion kinase (FAK) is a potential target for the treatment of primary cancers as well as prevention of tumor metastasis. To understand the structural and chemical features of FAK inhibitors, we report comparative molecular field analysis (CoMFA) for the series of 7H-pyrrolo(2,3-d)pyrimidines. The CoMFA models showed good correlation between the actual and predicted values for training set molecules. Our results indicated the ligand-based alignment has produced better statistical results for CoMFA ($q^2$ = 0.505, $r^2$ = 0.950). Both models were validated using test set compounds, and gave good predictive values of 0.537. The statistical parameters from the generated 3D-QSAR models were indicated that the data are well fitted and have high predictive ability. The contour map from 3D-QSAR models explains nicely the structure-activity relationships of FAK inhibitors and our results would give proper guidelines to further enhance the activity of novel inhibitors.

• Journal of Welding and Joining
• /
• v.15 no.6
• /
• pp.116-123
• /
• 1997

In GMA welding processes, bead contour and penetration patterns are criterion to estimate weld quality. Bead geometry is commonly defined with width, height and depth. When weaving is taken into account, selection of welding conditions is known to be difficult. Thus, empirical or trial-and-error method are usually introduced. This study examined the correlation of welding process variables including weaving parameters with bead geometry using srtificial neural networks(ANN). The main task of the Ann estimator is to realize the mapping characteristics from the sampled welding process variables to the actual bead geometry through training. After the neural network model is constructed, welding process variables for desired bead geometry is selected by inverse model. Experimental varification of the inverse model is conducted through actual welding.

• Korean Journal of Computational Design and Engineering
• /
• v.14 no.6
• /
• pp.382-389
• /
• 2009

Recently, knowledge management has been required in companies as a tool of competitiveness. Companies have constructed Enterprise Resource Planning(ERP) system in order to manage huge knowledge. But, it is not easy to formalize knowledge in organization. We focused on data mining system by genetic programming(GP). Data mining system by genetic programming can be useful tools to derive and extract the necessary information and knowledge from the huge accumulated data. However when we don't have enough amounts of data to perform the learning process of genetic programming, we have to reduce input parameter(s) or increase number of learning or training data. In this study, an enhanced data mining method combining Genetic Programming with Self organizing map, that reduces the number of input parameters, is suggested. Experiment results through a prototype implementation are also discussed.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2000.05a
• /
• pp.130-133
• /
• 2000

In this paper, a new design methodology named FNNN(Fuzzy Polynomial Neural Network) algorithm is proposed to identify the structure and parameters of fuzzy model using PNN(Polynomial Neural Network) structure and a fuzzy inference method. The PNN is the extended structure of the GMDH(Group Method of Data Handling), and uses several types of polynomials such as linear, quadratic and modified quadratic besides the biquadratic polynomial used in the GMDH. The premise of fuzzy inference rules defines by triangular and gaussian type membership function. The fuzzy inference method uses simplified and regression polynomial inference method which is based on the consequence of fuzzy rule expressed with a polynomial such as linear, quadratic and modified quadratic equation are used. Each node of the FPNN is defined as fuzzy rules and its structure is a kind of neuro-fuzzy architecture Several numerical example are used to evaluate the performance of out proposed model. Also we used the training data and testing data set to obtain a balance between the approximation and generalization of proposed model.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2002.05a
• /
• pp.134-138
• /
• 2002

The human faces do not have distinct features unlike other general objects. In general the features of eyes, nose and mouth which are first recognized when human being see the face are defined. These features have different characteristics depending on different human face. In this paper, We propose a face recognition algorithm using the hidden Markov model(HMM). In the preprocessing stage, we find edges of a face using the locally adaptive threshold scheme and extract features based on generic knowledge of a face, then construct a database with extracted features. In training stage, we generate HMM parameters for each person by using the forward-backward algorithm. In the recognition stage, we apply probability values calculated by the HMM to input data. Then the input face is recognized by the euclidean distance of face feature vector and the cross-correlation between the input image and the database image. Computer simulation shows that the proposed HMM algorithm gives higher recognition rate compared with conventional face recognition algorithms.