• Journal of the Korean Data and Information Science Society
• /
• v.22 no.5
• /
• pp.877-884
• /
• 2011

The decision tree is a representative algorithm of data mining and used in many domains such as retail target marketing, fraud detection, data reduction, variable screening, category merging, etc. This method is most useful in classification problems, and to make predictions for a target group after dividing it into several small groups. When we create a model of decision tree with a large number of input variables, we suffer difficulties in exploration and analysis of the model because of complex trees. And we can often find some association exist between input variables by external variables despite of no intrinsic association. In this paper, we study on the removal method of unnecessary input variables using multiple external association rules. And then we apply the removal method to actual data for its efficiencies.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.49 no.3
• /
• pp.145-156
• /
• 2000

In this paper, we propose a new methodology which includes the optimal design procedure of Polynomial Neural Networks(PNN) structure for model identification of complex and nonlinear system. The proposed PNN algorithm is based on GMDA(Group Method of Data handling) method and its structure is similar to Neural Networks. But the structure of PNN is not fixed like in conventional Neural Networks and can be generated. The each node of PNN structure uses several types of high-order polynomial such as linear, quadratic and cubic, and is connected as various kinds of multi-variable inputs. In other words, the PNN uses high-order polynomial as extended type besides quadratic polynomial used in GMDH, and the number of input of its node in each layer depends on that of variables used in the polynomial. The design procedure to obtain an optimal model structure utilizing PNN algorithm is shown in each stage. The study is illustrated with the aid of pH neutralization process data besides representative time series data for gas furnace process used widely for performance comparison, and shows that the proposed PNN algorithm can produce the model with higher accuracy than previous other works. And performance index related to approximation and prediction capabilities of model is evaluated and also discussed.

• Journal of the Korean Data and Information Science Society
• /
• v.23 no.2
• /
• pp.299-307
• /
• 2012

Data mining is a method of searching for an interesting relationship among items in a given database. The decision tree is a typical algorithm of data mining. The decision tree is the method that classifies or predicts a group as some subgroups. In general, when researchers create a decision tree model, the generated model can be complicated by the standard of model creation and the number of input variables. In particular, if the decision trees have a large number of input variables in a model, the generated models can be complex and difficult to analyze model. When creating the decision tree model, if there are marginally conditional variables (intervening variables, external variables) in the input variables, it is not directly relevant. In this study, we suggest the method of creating a decision tree using marginally conditional variables and apply to actual data to search for efficiency.

• Journal of Korean Neurosurgical Society
• /
• v.64 no.1
• /
• pp.13-22
• /
• 2021

Objective : High precision and accuracy are expected in gamma knife radiosurgery treatment. Because of the requirement of clinically applying complex radiation and dose gradients together with a rapid radiation decline, a dedicated quality assurance program is required to maintain the radiation dosimetry and geometric accuracy and to reduce all associated risk factors. This study investigates the validity of Leksell Gamma plan (LGP)10.1.1 system of 5th generation Gamma Knife Perfexion as modified variable ellipsoid modeling technique (VEMT) method. Methods : To verify LGP10.1.1 system, we compare the treatment plan program system of the Gamma Knife Perfexion, that is, the LGP, with the calculated value of the proposed modified VEMT program. To verify a modified VEMT method, we compare the distributions of the dose of Gamma Knife Perfexion measured by Gafchromic EBT3 and EBT-XD films. For verification, the center of an 80 mm radius solid water phantom is placed in the center of all sectors positioned at 16 mm, 4 mm and 8 mm; that is, the dose distribution is similar to the method used in the x, y, and z directions by the VEMT. The dose distribution in the axial direction is compared and analyzed based on Full-Width-of-Half-Maximum (FWHM) evaluation. Results : The dose profile distribution was evaluated by FWHM, and it showed an average difference of 0.104 mm for the LGP value and 0.130 mm for the EBT-XD film. Conclusion : The modified VEMT yielded consistent results in the two processes. The use of the modified VEMT as a verification tool can enable the system to stably test and operate the Gamma Knife Perfexion treatment planning system.

• Computational Structural Engineering
• /
• v.7 no.1
• /
• pp.91-98
• /
• 1994

Design sensitivity analysis method for the vibration of vehicle structure is developed using adjoint variable method. A variational approach with complex response method is used to derive sensitivity expression. To evaluate sensitivity, FEM analysis of ship deck and vehicle structure are performed using MSC/NASTRAN installed in the super computer CRAY2S, and sensitivity computation is performed by PC. The accuracy of sensitivity is verified by the results of finite difference method. When compared to structural analysis time on CRAY2S, sensitivity computation is remarkably economical. The sensitivity of vehicle frame can be used to reduce the vibration responses such as displacement and acceleration of vehicle.

• Journal of Electrical Engineering and Technology
• /
• v.2 no.2
• /
• pp.208-215
• /
• 2007

The significant drawback of the permanent magnet linear synchronous motor (PMLSM) is force ripples, which are generated by the distortion of the stator flux linkage distributions, cogging forces caused by the interaction of the permanent magnet and the iron core and the end effects. This will deteriorate the performance of the drive system in high precision applications. The PMLSM and its parasitic effects are analyzed and modeled using the complex state-variable approach. To minimize the force ripple and realize the high precision control, the components of force ripples are extracted first and then compensated by injecting the instantaneous current to counteract the force ripples. And this method of the PMLSM system is realized by the field oriented control method. In order to verify the validity of this proposed method, the system simulations are carried out and the results are analyzed. The effectiveness of the proposed force ripples reduction method can be seen according to the comparison between the compensation and non-compensation cases.

• Communications for Statistical Applications and Methods
• /
• v.29 no.1
• /
• pp.65-83
• /
• 2022

Although various statistical methods have been developed to map time-dependent genetic factors, most identified genetic variants can explain only a small portion of the estimated genetic variation in longitudinal traits. Gene-gene and gene-time/environment interactions are known to be important putative sources of the missing heritability. However, mapping epistatic gene-gene interactions is extremely difficult due to the very large parameter spaces for models containing such interactions. In this paper, we develop a Gaussian process (GP) based nonparametric Bayesian variable selection method for longitudinal data. It maps multiple genetic markers without restricting to pairwise interactions. Rather than modeling each main and interaction term explicitly, the GP model measures the importance of each marker, regardless of whether it is mostly due to a main effect or some interaction effect(s), via an unspecified function. To improve the flexibility of the GP model, we propose a novel grid-based method for the within-subject dependence structure. The proposed method can accurately approximate complex covariance structures. The dimension of the covariance matrix depends only on the number of fixed grid points although each subject may have different numbers of measurements at different time points. The deviance information criterion (DIC) and the Bayesian predictive information criterion (BPIC) are proposed for selecting an optimal number of grid points. To efficiently draw posterior samples, we combine a hybrid Monte Carlo method with a partially collapsed Gibbs (PCG) sampler. We apply the proposed GP model to a mouse dataset on age-related body weight.

• Structural Engineering and Mechanics
• /
• v.89 no.2
• /
• pp.199-211
• /
• 2024

The loading capacity of engineering structures/components reduces after the initiation and propagation of crack eventually leads to the final failure. Hence, it becomes essential to deal with the crack and its effects at the design and simulation stages itself, by detecting the prone area of the fracture. The phase-field (PF) method has been accepted widely in simulating fracture problems in complex geometries. However, most of the PF methods are formulated with second order continuity theoryinvolving C⁰ continuity. In the present study, PF method based on fourth-order (i.e., higher order) theory, maintaining C¹ continuity has been proposed for ductile fracture simulation. The formulation includes fourth-order derivative terms of phase field variable, varying between 0 and 1. Applications of fourth-order PF theory to ductile fracture simulation resulted in novelty in this area. The proposed formulation is numerically solved using a two-dimensional finite element (FE) framework in 3-layered manner system. The solutions thus obtained from the proposed fourth order theory for different benchmark problems portray the improvement in the accuracy of the numerical results and are well matched with experimental results available in the literature. These results are also compared with second-order PF theory and a comparison study demonstrated the robustness of the proposed model in capturing ductile behaviour close to experimental observations.

• Journal of the Korean Institute of Telematics and Electronics T
• /
• v.35T no.3
• /
• pp.15-24
• /
• 1998

In this paper, we present new method of fractal image coding based on iterated function system(IFS) suggested by Barnsley. In previous Fractal coding using full searching algorithm, the quality of reconstructed image was better than other fractal coding method's, but it took a long time in that algorithm for searching domain blocks matched. And it is performed through linear affine transform, therefore it is difficult to approximate the complex range blocks. In this paper, using quadtree partitioning, complex blocks are divided into more smaller blocks, and simple blocks are merged to more larger blocks. So, we can got more precisely approximated range blocks and reduce the number of transformations. Hence, we have improved the compression ratio. In addition, we restrict the region of searching domains in order to reduce the searching time and coding time. Compared with full searching algorithm, we reduced coding time drastically, and quality of reconstructed image was better in terms subjective criteria. And compared with Monro's, our method is slower, but we could obtain a reconstructed image with better quality.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.46 no.6
• /
• pp.95-101
• /
• 2009

Among the adaptive noise cancellers (ANC), the least mean square (LMS) algorithm has probably become the most popular algorithm because of its robustness, good tracking properties, and simplicity of implementation. However, it has non-uniform convergence and a trade-off between the rate of convergence and excess mean square error (EMSE). To overcome these shortcomings, a number of variable step size least mean square (VSSLMS) algorithms have been researched for years. These LMS algorithms use a complex variable step method approach for rapid convergence but need high computational complexity. A variable step approach can impair the simplicity and robustness of the LMS algorithm. The proposed instantaneous gain control (IGC) algorithm uses the instantaneous gain value of the original signal and the noise signal. As a result, the IGC algorithm can reduce computational complexity and maintain better performance.