• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.8 no.5
• /
• pp.384-393
• /
• 2015

This paper proposes an optimized classifier based on feature elimination (OCFE) for gene selection with combining two feature elimination methods, ReliefF and SVM-RFE. ReliefF algorithm is filter feature selection which rank the data by the importance of the data. SVM-RFE algorithm is a wrapper feature selection which wrapped the data and rank the data based on the weight of feature. With combining these two methods we get less error rate average, 0.3016138 for OCFE and 0.3096779 for SVM-RFE. The proposed method also get better accuracy with 70% for OCFE and 69% for SVM-RFE.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2012.10a
• /
• pp.143-146
• /
• 2012

Enforcement of private information preservation law the infringement accident which is still caused by despite with private information outflow occurs. The private information outflow avoids finance and to become the fire tube. Analyzes the pattern of private information from search of private information and detection process and the research which it extracts and the research is necessary in about perfection elimination of the private information which is unnecessary. From the research which it sees it researched a pattern extraction research and a complete elimination method in about private information protection and it did the pattern extraction and a complete elimination experiment of private information.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2013.05a
• /
• pp.213-216
• /
• 2013

Big data, the use of privacy has been increasing to the development of wireless network infrastructure or technology development and wired Internet. By the way, Enforcement of private data preservation law the infringement accident which is still caused by despite with private data outflow occurs. The private data outflow avoids finance and to become the fire tube. Analyzes the pattern of private data from search of private data and detection process and the research which it extracts and the research is necessary in about perfection elimination of the private data which is unnecessary. From the research which it sees it researched a pattern extraction research and a complete elimination method in about private data protection and it did the pattern extraction and a complete elimination experiment of private data.

• The Korean Journal of Applied Statistics
• /
• v.20 no.3
• /
• pp.531-540
• /
• 2007

The recursive feature elimination for support vector machine is known to be useful in selecting relevant genes. Since the criterion for choosing relevant genes is the absolute value of a coefficient, the recursive feature elimination may suffer from a scaling problem. We propose a modified version of the recursive feature elimination algorithm using bootstrap. In our method, the criterion for determining relevant genes is the absolute value of a coefficient divided by its standard error, which accounts for statistical variability of the coefficient. Through numerical examples, we illustrate that our method is effective in gene selection.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.32 no.2A
• /
• pp.162-168
• /
• 2007

This paper proposes two detection algorithms for Multi-user Space Time Block Code systems. The first one is linear detection Gaussian Elimination algorithm, and then it combined with Ordered Successive Cancellation to get better performance. The comparisons between receiver and other popular receivers, including linear receivers are provided. It will be shown that the performance of Gaussian Elimination receiver is similar but more simplicity than linear detection algorithms and performance of Gaussian Elimination Ordered Successive Cancellation superior as compared to other linear detection method.

• Journal of Electrical Engineering and Technology
• /
• v.6 no.3
• /
• pp.334-341
• /
• 2011

This paper presents accurate solutions for nonlinear transcendental equations of the selective harmonic elimination technique used in three-phase PWM inverters feeding the induction motor by particle swarm optimization (PSO). With the proposed approach, the required switching angles are computed efficiently to eliminate low order harmonics up to the $23^{rd}$ from the inverter voltage waveform, whereas the magnitude of the fundamental component is controlled to the desired value. A set of solutions and the evaluation of the proposed method are presented. The obtained results prove that the algorithm converges to a precise solution after several iterations. The salient contribution of the paper is the application of the particle swarm algorithm to attenuate successfully any undesired loworder harmonics from the inverter output voltage. The current paper demonstrates that the PSO is a promising approach to control the operation of a three-phase voltage source inverter with a selective harmonic elimination strategy to be applied in induction motor drives.

• Journal of Power Electronics
• /
• v.17 no.4
• /
• pp.914-922
• /
• 2017

Multilevel converters have attracted a lot of attention in recent years. The efficiency parameters of a multilevel converter such as the switching losses and total harmonic distortion (THD) mainly depend on the modulation strategy used to control the converter. Among all of the modulation techniques, the selective harmonic elimination (SHE) method is particularly suitable for high-power applications due to its low switching frequency and high quality output voltage. This paper proposes a new expression for the SHE problem in five-level converters. Based on this new expression, a simple analytical method is introduced to determine the feasible modulation index intervals and to calculate the exact value of the switching angles. For each selected harmonic, this method presents three-level or five-level waveforms according to the value of the modulation index. Furthermore, a flowchart is proposed for the real-time implementation of this analytical method, which can be performed by a simple processor and without the need of any lookup table. The performance of the proposed algorithm is evaluated with several simulation and experimental results for a single phase five-level diode-clamped inverter.

• Proceedings of the KIEE Conference
• /
• 1990.11a
• /
• pp.129-131
• /
• 1990

When insulating oil is flowed through a pipe, the electrostatic charge in the oil will be increased. Using inflammability oil, an electrostatic phenomena which may cause significantly problems as explosion must be circumvented by some methods. In this paper we made a new electrode which was applied AC field to the flowing oil for reducing electrostatic charge. Before we developed the new method of an electrostatic elimination, the electrification experiment was performed by using the same electrode which was supplied DC voltage to realize electrified conditions. As a result of this study, the effect of the electrostatic elimination could be observed.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.65 no.10
• /
• pp.1774-1779
• /
• 2016

This paper proposes a portable cable fault detection system with automatic fault distinction and distance measurement using time-frequency correlation and reference signal elimination method and automatic fault classification algorithm in order to have more accurate fault determination and location detection than conventional time domain refelectometry (TDR) system despite increased signal attenuation due to the long distance to cable fault location. The performance of the developed system method was validated via an experiment in the test field constructed for the standardized performance test of power cable fault location equipments. The performance evaluation showed that accuracy of the developed system is less than 1.34%. Also, an error of automatic fault type and location by detection of phase and peak value through elimination of the reference signal and normalization of correlation coefficient and automatic fault classification algorithm not occurred.

• Current Optics and Photonics
• /
• v.3 no.6
• /
• pp.496-501
• /
• 2019

We propose a method to eliminate the noise of a disordered medium optical transmission matrix. Gaussian noise exists whenever light passes through the medium, during the measurement of the transmission matrix and thus cannot be ignored. Experiments and comparison of noise eliminating before and after are performed to illustrate the effectiveness and advance presented by our method. After noise elimination, the results of focusing and imaging are better than the effect before noise elimination, and the measurement of the transmission matrix is more consistent with the theoretical analysis as well.