• The Korean Journal of Applied Statistics
• /
• v.18 no.1
• /
• pp.137-146
• /
• 2005

Neural network is one of the most notable technique. It usually provides more powerful forecasting models than the traditional time series techniques. Employing the Ensemble technique in forecasting model, one should provide a initial distribution. Usually the uniform distribution is assumed so that the initialization is noninformative. However, it would be expected a sequential informative initialization based on data rather than the uniform initialization gives further reduction in forecasting error. In this note, a modified Ensemble algorithm using sequential initial probability is developed. The sequential distribution is designed to have much weight on the recent data.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.18 no.3
• /
• pp.77-82
• /
• 2018

This paper is concerned with the performance analysis of a modified stop-and-go generalized multi-modulus algorithm (MSAG-GMMA) adaptive blind equalization algorithm with variable step size. The proposed algorithm multiplies the fixed step size by the error signal of the decision-oriented algorithm in the equalization coefficient update equation, and changes the step size according to the error size. Also, the MSAG-GMMA having a fixed step size is operated so as to maintain a fast convergence speed from a certain threshold to a steady state by determining the error signal size of the decision-directed algorithm, and when the MSAG-GMMA to work To evaluate the performance of the proposed algorithm, we use the ensemble ISI, ensemble-averaged MSE, and equalized constellation obtained from the output of the equalizer as the performance index. Simulation results show that the proposed algorithm has faster convergence speeds than MMA, GMMA, and MSAG-GMMA and has a small residual error in steady state.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.11 no.5
• /
• pp.681-688
• /
• 2000

In this paper, a planar antenna that is small and light, is designed and analyzed aiming handset antenna of IMT-2000. Employing the Ensemble simulator based on a MoM, design-parameters are found to determine a resonant frequency. Therefore, it is analyzed with the Ensemble simulation and FDTD numerical for resonating at the allocated frequency for IMT-2000 in the fixed antenna dimension of 21$^{\circ}$wing angle that is a design parameter. Analyzing with FDTD method, Though the results of FDTD are very exact, this analysis introduces errors due to the staircasing approximation in the slope of bowtie. To reduce this error, it is divided to 4-ranges where the cell contains the boundary of perfect conductor/free space. Then, each range is calculated by different by different equation, which modify the H-field to add the component of the area and length of the cell filled with free space. Therefore, the modified FDTD algorithm provided with a narrow bandwidth of return loss calculated with a standard FDTD algorithm that can be extended to the desired ranges.

• Electrical & Electronic Materials
• /
• v.10 no.10
• /
• pp.1000-1004
• /
• 1997

In this paper the current-voltage characteristics of a submicron GaAs MESFET is simulated by using the self-consistent ensemble Monte Carlo method. The numerical algorithm employed in solving the two-dimensional Poisson equation is the successive over-relaxation(SOR) method. The total number of employed superparticles is about 1000 and the field adjusting time is 10fs. To obtain the steady-state results the simulation is performed for 10ps at each bias condition. The simulation results show the average electron velocity is modified by the gate voltage.

• East Asian Economic Review
• /
• v.25 no.4
• /
• pp.403-424
• /
• 2021

AdaBoost tweaks the sample weight for each training set used in the iterative process, however, it is demonstrated that it provides more correlated errors as the boosting iteration proceeds if models' accuracy is high enough. Therefore, in this study, we propose a novel way to improve the performance of the existing AdaBoost algorithm by employing heterogeneous models and a stochastic twist. By employing the heterogeneous ensemble, it ensures different models that have a different initial assumption about the data are used to improve on diversity. Also, by using a stochastic algorithm with a decaying convergence rate, the model is designed to balance out the trade-off between model prediction performance and model convergence. The result showed that the stochastic algorithm with decaying convergence rate's did have a improving effect and outperformed other existing boosting techniques.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.19 no.2
• /
• pp.279-286
• /
• 2015

EMD is a fully data-driven signal processing method without using any predetermined basis function and requiring any user parameters setting. However EMD experiences a problem of mode mixing which interferes with decomposing the signal into similar oscillations within a mode. To overcome the problem, EEMD method was introduced. The algorithm performs the EMD method over an ensemble of the signal added independent identically distributed white noise of the same standard deviation. Even so EEMD created problems when the decomposition is complete. The ensemble of different signal with added noise may produce different number of modes and the reconstructed signal includes residual noise. This paper propose an modified EEMD method to overcome mode mixing of EMD, to provide an exact reconstruction of the original signal, and to separate modes with lower cost than EEMD's. The experimental results show that the proposed method provides a better separation of the modes with less number of sifting iterations, costs 20.87% for a complete decomposition of the signal and demonstrates superior performance in the signal reconstruction, compared with EEMD.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.51 no.4
• /
• pp.152-159
• /
• 2002

Fitness sharing that shares fitness if calculated distance between individuals is smaller than sharing radius is one of the representative speciation methods and can complement evolutionary algorithm which converges one solution. Recently, there are many researches on designing neural network architecture using evolutionary algorithm but most of them use only the fittest solution in the last generation. In this paper, we elaborate generating diverse neural networks using fitness sharing and combing them to compute outputs then, propose calculating distance between individuals using modified Kullback-Leibler entropy for improvement of fitness sharing performance. In the experiment of Australian credit card assessment, breast cancer, and diabetes in UCI database, proposed method performs better than not only simple average output or Pearson Correlation but also previous published methods.

• Asia pacific journal of information systems
• /
• v.29 no.4
• /
• pp.789-816
• /
• 2019

Text classification is a challenging task, especially when dealing with a huge amount of text data. The performance of a classification model can be varied depending on what type of words contained in the document corpus and what type of features generated for classification. Aside from proposing a new modified version of the existing algorithm or creating a new algorithm, we attempt to modify the use of data. The classifier performance is usually affected by the quality of learning data as the classifier is built based on these training data. We assume that the data from different domains might have different characteristics of noise, which can be utilized in the process of learning the classifier. Therefore, we attempt to enhance the robustness of the classifier by injecting the heterogeneous data artificially into the learning process in order to improve the classification accuracy. Semi-supervised approach was applied for utilizing the heterogeneous data in the process of learning the document classifier. However, the performance of document classifier might be degraded by the unlabeled data. Therefore, we further proposed an algorithm to extract only the documents that contribute to the accuracy improvement of the classifier.

• Geomechanics and Engineering
• /
• v.32 no.6
• /
• pp.583-600
• /
• 2023

The resilient modulus (M_R) of various pavement materials plays a significant role in the pavement design by a mechanistic-empirical method. The M_R determination is done by experimental tests that need time and money, along with special experimental tools. The present paper suggested a novel hybridized extreme gradient boosting (XGB) structure for forecasting the M_R of modified base materials subject to wet-dry cycles. The models were created by various combinations of input variables called deep learning. Input variables consist of the number of W-D cycles (WDC), the ratio of free lime to SAF (CSAFR), the ratio of maximum dry density to the optimum moisture content (DMR), confining pressure (σ₃), and deviatoric stress (σ_d). Two XGB structures were produced for the estimation aims, where determinative variables were optimized by particle swarm optimization (PSO) and black widow optimization algorithm (BWOA). According to the results' description and outputs of Taylor diagram, M1 model with the combination of WDC, CSAFR, DMR, σ₃, and σ_d is recognized as the most suitable model, with R² and RMSE values of BWOA-XGB for model M1 equal to 0.9991 and 55.19 MPa, respectively. Interestingly, the lowest value of RMSE for literature was at 116.94 MPa, while this study could gain the extremely lower RMSE owned by BWOA-XGB model at 55.198 MPa. At last, the explanations indicate the BWO algorithm's capability in determining the optimal value of XGB determinative parameters in M_R prediction procedure.