• The KIPS Transactions:PartB
• /
• v.17B no.5
• /
• pp.381-388
• /
• 2010

Naive Bayesian learning has been widely used in many data mining applications, and it performs surprisingly well on many applications. However, due to the assumption that all attributes are equally important in naive Bayesian learning, the posterior probabilities estimated by naive Bayesian are sometimes poor. In this paper, we propose more fine-grained weighting methods, called value weighting, in the context of naive Bayesian learning. While the current weighting methods assign a weight to each attribute, we assign a weight to each attribute value. We investigate how the proposed value weighting effects the performance of naive Bayesian learning. We develop new methods, using gradient descent method, for both value weighting and feature weighting in the context of naive Bayesian. The performance of the proposed methods has been compared with the attribute weighting method and general Naive bayesian, and the value weighting method showed better in most cases.

• Journal of Broadcast Engineering
• /
• v.21 no.6
• /
• pp.889-898
• /
• 2016

Virtual Bass System (VBS) is widely used to extend the lower frequency limit of small loudspeakers, which generates harmonics of a fundamental frequency. The perceptual quality of the VBS is highly dependent on the harmonic weighting strategy. There have been several weighting methods, including exponential attenuation and timbre matching. However, it is essential to match phases between harmonics in the original signal and generate harmonics to precisely convey the weighting strategy. This paper shows the limitations of the previous harmonic weighting schemes and proposes a new harmonic weighting scheme. The proposed weighting scheme proposes phase matching between the original and generated harmonics and varies the slope of the attenuation weighting dynamically according to the missing fundamental frequency. Objective and subjective tests show that the proposed harmonic weighting scheme provides more natural and effective bass perception in a limited situation than the conventional schemes, which implies that the phase matching is essential for the high quality bass enhancement.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.22 no.1
• /
• pp.41-45
• /
• 1985

A degradation of the signal-to-noise ratio (SNR) in a digital moving target indicator system (DMTI System) can be Improved by weighting of the DMTI output pulses before the integration process. and it Is shown that the signal-to-noise ratio Is unproved as the nunlber of delay line cancellers. It is Known that signal-to-noise ratio obtained with optimum weighting is greater than that obtained with binomial weighting. An improvement performance of the signal-to-noise ratio for each case is presented through the results or a computter simulation.

• Journal of Korea Water Resources Association
• /
• v.36 no.5
• /
• pp.751-760
• /
• 2003

Error analysis of finite difference equation on the Muskingum-Cunge flood routing method with free time and space weighting factor was carried out. The error analysis shows that the numerical solution of the Muskingum-Cunge method becomes diverged with time when the sum of time weighting factor and space weighting factor is greater than 1.0. Numerical diffusion increases when the sum of time weighting factor and space weighting factor decreases. Numerical diffusion and numerical oscillation increase when the grid resolution is coarse. Numerical experiments and field applications show that the Muskingum-Cunge method with free space weighting factor is more effective for simulating the flood routing with great peak diminution than conventional Muskingum-Cunge method with fixed space weighting factor, 0.5.

• Journal of Institute of Control, Robotics and Systems
• /
• v.12 no.7
• /
• pp.631-636
• /
• 2006

In this paper, an algorithm of time series function forecasting using GMDH(group method of data handling) algorithm that gives more weight to the recent data is proposed. Traditional methods of GMDH forecasting gives same weights to the old and recent data, but by the point of view that the recent data is more important than the old data to forecast the future, an algorithm that makes the recent data contribute more to training is proposed for more accurate forecasting. The average error rate of electric power demand forecasting by the traditional GMDH algorithm which does not use data weighting algorithm is 0.9862 %, but as the result of applying the data weighting GMDH algorithm proposed in this paper to electric power forecasting demand the average error rate by the algorithm which uses data weighting algorithm and chooses the best data weighting rate is 0.688 %. Accordingly in forecasting the electric power demand by GMDH the proposed method can acquire the reduced error rate of 30.2 % compared to the traditional method.

• Journal of the Korean Society for Precision Engineering
• /
• v.29 no.9
• /
• pp.986-994
• /
• 2012

This paper is concerned with the application of the vision control algorithm with weighting matrix in robot point placement task. The proposed vision control algorithm involves four models, which are the robot kinematic model, vision system model, the parameter estimation scheme and robot joint angle estimation scheme. This proposed algorithm is to make the robot move actively, even if relative position between camera and robot, and camera's focal length are unknown. The parameter estimation scheme and joint angle estimation scheme in this proposed algorithm have form of nonlinear equation. In particular, the joint angle estimation model includes several restrictive conditions. For this study, the weighting matrix which gave various weighting near the target was applied to the parameter estimation scheme. Then, this study is to investigate how this change of the weighting matrix will affect the presented vision control algorithm. Finally, the effect of the weighting matrix of robot vision control algorithm is demonstrated experimentally by performing the robot point placement.

• Journal of the Architectural Institute of Korea Planning & Design
• /
• v.34 no.12
• /
• pp.49-54
• /
• 2018

Applying metric(distance) factor as weighting to spatial syntax is known to not contribute to the explanatory power for the human movement behavior as compared to the geometric(angle) factor according to the negative results of several related studies. However, Kim & Piao (2017) assumed that there is not a problem of the metric factor itself but a problem of the way of applying the metric factor as weighting, and presented a new possibility of the metric factor as weighting by proposing and verifying the methods of applying the metric weighting, which are different from the existing ones. The purpose of this study is to propose advanced methods of applying the metric weighting to space syntax, and to verify whether they contribute to the improvement of explanatory power of space syntax analysis. In this paper, we propose functions for combined depth of distance-step that combine the distance-weighted depth function with the step depth function and apply them to axial segment analysis to check the improvement of explanatory power of them.

• Journal of applied mathematics & informatics
• /
• v.30 no.3_4
• /
• pp.593-602
• /
• 2012

In this paper, we first introduce a special type of multisplitting method with different weighting scheme, and then we provide convergence results of multisplitting methods with different weighting schemes corresponding to both the AOR-like multisplitting and the SSOR-like multisplitting.

• Proceedings of the Earthquake Engineering Society of Korea Conference
• /
• 2002.03a
• /
• pp.475-482
• /
• 2002

This paper provides the systematic procedure to determine the weighting matrices of optimal controllers considering structural energy. Optimal controllers consist of LQR and ILQR. The weighting matrices are needed first in the conventional optimal control design strategy. However, they are in general dependent on the experienced knowledge of controll designers. Applying the Lyapunov function to the total structural energy and using the contrition that its derivative is negative, we can determine the weighting matrices without difficulty. It is proven that the control efficiency is achieved well for LQR and ILQR.

• Proceedings of the IEEK Conference
• /
• 1998.10a
• /
• pp.1267-1270
• /
• 1998

This paper analyzes the characteristics of wide-band one-shot beam formed by using all sensors of array at once, as variation of weighting width. Gaussian function is applied to each sensor as a role of weighting. As the results of the simulation for nested linear array having 17 sensors for each octave, as the width goes wider the directivity index(DI) becomes lower but more even and the variation of beamwidth becomes smaller. It is confirmed, therefore, that weighting width is carefully decided in consideration of DI level, DI stability and the beamwidth.