• Smart Structures and Systems
• /
• 제11권1호
• /
• pp.123-134
• /
• 2013

Identifying structural modal parameters, especially those modes within high frequency range, from ambient data is still a challenging problem due to various kinds of uncertainty involved in vibration measurements. A procedure applying an ensemble empirical mode decomposition (EEMD) method is proposed for accurate and robust structural modal identification. In the proposed method, the EEMD process is first implemented to decompose the original ambient data to a set of intrinsic mode functions (IMFs), which are zero-mean time series with energy in narrow frequency bands. Subsequently, a Sub-PolyMAX method is performed in narrow frequency bands by using IMFs as primary data for structural modal identification. The merit of the proposed method is that it performs structural identification in narrow frequency bands (take IMFs as primary data), unlike the traditional method in the whole frequency space (take original measurements as primary data), thus it produces more accurate identification results. A numerical example and a multiple-span continuous steel bridge have been investigated to verify the effectiveness of the proposed method.

• 한국언어정보학회지:언어와정보
• /
• 제10권1호
• /
• pp.21-46
• /
• 2006

This study proposes a Word Sense Disambiguation (WSD) algorithm, based on concept learning with special emphasis on statistically meaningful lowest frequency words. Previous works on WSD typically make use of frequency of collocation and its probability. Such probability based WSD approaches tend to ignore the lowest frequency words which could be meaningful in the context. In this paper, we show an algorithm to extract and make use of the meaningful lowest frequency words in WSD. Learning method is adopted from the Find-Specific algorithm of Mitchell (1997), according to which the search proceeds from the specific predefined hypothetical spaces to the general ones. In our model, this algorithm is used to find contexts with the most specific classifiers and then moves to the more general ones. We build up small seed data and apply those data to the relatively large test data. Following the algorithm in Yarowsky (1995), the classified test data are exhaustively included in the seed data, thus expanding the seed data. However, this might result in lots of noise in the seed data. Thus we introduce the 'maximum a posterior hypothesis' based on the Bayes' assumption to validate the noise status of the new seed data. We use the Naive Bayes Classifier and prove that the application of Find-Specific algorithm enhances the correctness of WSD.

• 한국정보시스템학회지:정보시스템연구
• /
• 제27권3호
• /
• pp.123-140
• /
• 2018

Purpose The purpose of this study is to explore the optimal trading frequency which is useful for stock price prediction by using deep learning for charting image data. We also want to identify the appropriate time for accurate forecasting of stock price when performing pattern analysis. Design/methodology/approach In order to find the optimal trading frequency patterns and forecast timings, this study is performed as follows. First, stock price data is collected using OpenAPI provided by Daishin Securities, and candle chart images are created by data frequency and forecasting time. Second, the patterns are generated by the charting images and the learning is performed using the CNN. Finally, we find the optimal trading frequency patterns and forecasting timings. Findings According to the experiment results, this study confirmed that when the 10 minute frequency data is judged to be a decline pattern at previous 1 tick, the accuracy of predicting the market frequency pattern at which the market decreasing is 76%, which is determined by the optimal frequency pattern. In addition, we confirmed that forecasting of the sales frequency pattern at previous 1 tick shows higher accuracy than previous 2 tick and 3 tick.

• 한국지진공학회:학술대회논문집
• /
• 한국지진공학회 2002년도 추계 학술발표회 논문집
• /
• pp.35-42
• /
• 2002

Frequency responses for most of the local seismic sensors in Korea have been roughly checked by mutual comparison of Fourier spectra of seismic records from accelerometer and seismometer, both of which are installed at the same location. Especially, because the frequency content of the seismic energy is usually above 1 Hz for local earthquakes, the reliability of low frequency response could have not been evaluated. Fortunately a recent large earthquake, Ms=7.2 on 02/06/29 containing dominant low frequency energy makes it possible to check the low frequency response of the seismic sensors, especially EpiSensor and JC-V100. Considering two types of sensor pairs, (STS-2 and EpiSensor, JC-V100 and EpiSensor), the low frequency response of EpiSensor is confirmed first by comparison with STS-2 which has proved low frequency response. Second, reliable low frequency limit of instrumentally corrected seismic data from JC-V100 data is estimated to be about 0.03 Hz by comparison with EpiSensor data.

• 한국가스학회지
• /
• 제18권3호
• /
• pp.60-66
• /
• 2014

본 연구에서는 국내 매설 고압가스배관의 사고빈도 데이터 구축 시 활용할 수 있도록 국외에서 발표하고 있는 고압가스배관 사고빈도 데이터에 대해 고찰 하였다. 고압가스배관 사고빈도 데이터의 대표적인 것으로는 미국의 DOT, 유럽의 EGIG 및 영국의 UKOPA가 있다. 국외 사고빈도 데이터의 국내 적용 가능성을 확인하기 위하여 이들을 비교 분석한 결과 EGIG 데이터가 국내 실정에 더 적합하였다. EGIG 8차 보고서의 사고빈도 데이터를 사용하여 비선형회귀분석을 수행한 결과 배관 설치 연도에 따른 지수형의 곡선을 얻었다. 향후 전체 사고빈도의 약 50%를 차지하고 있는 타공사 부분과 국내 데이터와 국외 데이터의 합성에 대한 집중적인 연구가 필요하다.

• 정보처리학회논문지:컴퓨터 및 통신 시스템
• /
• 제4권9호
• /
• pp.321-326
• /
• 2015

본 논문에서는 가청 주파수 영역 중 사람들에게 거의 들리지 않는 고주파와 순환 중복 검사 기법을 이용하여 스마트 기기 간의 신뢰성 있는 데이터를 무선으로 전송하는 알고리즘을 제안한다. 제안 알고리즘은 스마트 기기의 내장 스피커에서 출력할 수 있는 가청 주파수 영역(20 Hz~22 kHz) 중 고주파 영역인 18 kHz~22 kHz를 사용한다. 이때 데이터의 전송량을 높이기 위해 고주파 영역 1(18.0 kHz~21.2 kHz)에서 여러 개의 주파수를 혼합하여 전달하며, 이와 동시에 전송 데이터의 정확성을 높이기 위해 고주파 영역 2(21.2 kHz~22.0 kHz)에서 순환 중복 검사를 위한 체크섬을 전달한다. 제안 방법의 데이터 전송 속도와 정확성을 확인하기 위해 스마트 북과 스마트 기기 간에 데이터 전달 실험을 하였다. 그 결과 평균 235 ms에 32 bits 데이터를 전송할 수 있었으며, 전송 성공률은 99.47%, 그리고 순환 중복 검사에 의한 에러 검출률은 0.53%인 것을 확인하였다. 따라서 제안 방법은 스마트 기기 간에 무선으로 데이터를 전송할 수 있는 유용한 기술이 될 것이다.

• Journal of Electrical Engineering and Technology
• /
• 제5권1호
• /
• pp.28-35
• /
• 2010

This paper presents new Adaptive Linear Combination Structure (ADALINE) for tracking/estimating voltage-current phasor and frequency of power system. To estimate the phasors and frequency from sampled data, the algorithm assumes that orthogonal coefficients and speed of angular frequency of power system are unknown parameters. With adequate sampled data, the estimation problem can be considered as a linear weighted least squares (LMS) problem. In addition to determining the phasors (orthogonal coefficients), the procedure estimates the power system frequency. The main algorithm is verified through a computer simulation and data from field. The proposed algorithm is tested with transient and dynamic behaviors during power swing, a step change of frequency upon islanding of small generators and disconnection of load. The algorithm shows a very high accuracy, robustness, fast response time and adaptive performance over a wide range of frequency, from 10 to 2000 Hz.

• Smart Structures and Systems
• /
• 제14권6호
• /
• pp.1197-1220
• /
• 2014

Accurate actuator tracking is critical to achieve reliable real-time hybrid simulation results for earthquake engineering research. The frequency-domain evaluation approach provides an innovative way for more quantitative post-simulation evaluation of actuator tracking errors compared with existing time domain based techniques. Utilizing the Fast Fourier Transform the approach analyzes the actuator error in terms of amplitude and phrase errors. Existing application of the approach requires using the complete length of the experimental data. To improve the computational efficiency, two techniques including data decimation and frequency decimation are analyzed to reduce the amount of data involved in the frequency-domain evaluation. The presented study aims to enhance the computational efficiency of the approach in order to utilize it for future on-line actuator tracking evaluation. Both computational simulation and laboratory experimental results are analyzed and recommendations on the two decimation factors are provided based on the findings from this study.

• 한국공작기계학회:학술대회논문집
• /
• 한국공작기계학회 2003년도 추계학술대회
• /
• pp.137-142
• /
• 2003

A fundamental study for developing a system of fault diagnosis of a pump is performed by using neural network. The acoustic signals were obtained and converted to frequency domain for normal products and artificially deformed products. The signals were obtained in various driving frequencies in order to obtain many types of data from a limited number of pumps. The acoustic data in frequency domain were managed to multiples of real driving frequency with the aim of easy comparison. The neural network model used in this study was 3-layer type composed of input, hidden, and output layer. The normalized amplitudes at the multiples of real driving frequency were chosen as units of input layer, Various sets of teach signals made from original data by eliminating some random cases were used in the training. The average errors were approximately proportional to the number of untaught data. The results showed neural network trained by acoustic signals can be used as a simple method far a detection of machine malfunction or fault diagnosis.

• 한국공작기계학회논문집
• /
• 제13권5호
• /
• pp.81-86
• /
• 2004

A fundamental study for developing a system of fault diagnosis of a pump is performed by using neural network. The acoustic signals were obtained and converted to frequency domain for normal products and artificially deformed products. The signals were obtained in various driving frequencies in order to obtain many types of data from a limited number of pumps. The acoustic data in frequency domain were managed to multiples of real driving frequency with the aim of easy comparison. The neural network model used in this study was 3-layer type composed of input, hidden, and output layer. The normalized amplitudes at the multiples of real driving frequency were chosen as units of input layer. Various sets of teach signals made from original data by eliminating some random cases were used in the training. The average errors were approximately proportional to the number of untaught data. The results showed neural network trained by acoustic signals can be used as a simple method for a detection of machine malfuction or fault diagnosis.