• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1999년도 제14차 학술회의논문집
• /
• pp.226-229
• /
• 1999

We have been developing a method to build models for time series using Genetic Programming. The proposed method has been applied to various kinds of time series e.g. computer-generated chaos, natural phenomena, and financial market indices etc. Now we apply the prediction method to time histories of seismic ground motion i.e. one-step-ahead prediction of seismographic amplitude. Waves of earthquakes are composed of P-waves and S-waves. They propagate in different speeds and have different characteristics. It is believed that P-waves arrive firstly and S-waves arrive secondly. Simulations were performed based on real data of Hyuganada earthquake which broke out at southern part of Kyushuu Island in Japan. To our surprise, prediction model built using the earthquake waves in early time can enough precisely predict main huge waves in later time. Lots of experiments lead us to conclude that every slice of data involves P-wave and S-wave. The simulation results suggest the GP-based prediction method can be utilized in alarm systems or dispatch systems in an emergency.

• 무역학회지
• /
• 제46권6호
• /
• pp.191-209
• /
• 2021

In this paper, various time series analysis models and machine learning models are presented for long-term prediction of export growth rate, and the prediction performance is compared and reviewed by RMSE and MAE. Export growth rate is one of the major economic indicators to evaluate the economic status. And It is also used to predict economic forecast. The export growth rate may have a negative (-) value as well as a positive (+) value. Therefore, Instead of using the ReLU function, which is often used for time series prediction of deep learning models, the PReLU function, which can have a negative (-) value as an output value, was used as the activation function of deep learning models. The time series prediction performance of each model for three types of data was compared and reviewed. The forecast data of long-term prediction of export growth rate was deduced by three forecast methods such as a fixed forecast method, a recursive forecast method and a rolling forecast method. As a result of the forecast, the traditional time series analysis model, ARDL, showed excellent performance, but as the time period of learning data increases, the performance of machine learning models including LSTM was relatively improved.

• 한국컴퓨터정보학회논문지
• /
• 제21권5호
• /
• pp.91-98
• /
• 2016

We propose a method to train a model that can predict the probability of a crime being committed. CCTV data by matching criminal events are required to train the crime prediction model. However, collecting CCTV data appropriate for training is difficult. Thus, we collected actual criminal records and converted them to an appropriate format using variables by considering a crime prediction environment and the availability of real-time data collection from CCTV. In addition, we identified new specific crime types according to the characteristics of criminal events and trained and tested the prediction model by applying neural network partial least squares for each crime type. Results show a level of predictive accuracy sufficiently significant to demonstrate the applicability of CCTV to real-time crime prediction.

• 한국지능시스템학회논문지
• /
• 제16권6호
• /
• pp.730-735
• /
• 2006

본 논문은 퍼지 시스템과 유전자 알고리즘을 이용하여 단기 전력 부하 예측 방법을 제안한다. 우선 유전자 알고리즘을 이용하여 최적의 퍼지 소속함수를 구한다. 최적의 퍼지 규칙과 시계열 입력 차이를 이용하여 보다 더 나은 예측 시스템을 구한다. 제안된 방법을 이용하여 단기 전력 부하 예측에서 좋은 결과를 얻었다. 또한 제안된 알고리즘에 대한 그래픽 사용자 인터페이스를 구현한다. 마지막으로, 전력부하에 대한 지역 예측 시스템을 구현한다.

• 한국항해항만학회지
• /
• 제35권9호
• /
• pp.735-740
• /
• 2011

It is necessary to develop the one stop system in order to protect our marine environment rapidly from oil spill accident. The purpose of this study is to develop real time database for oil spill prediction modeling and implement real time prediction modelling with ESI and server-client GIS based user interface. The existing oil spill prediction model cannot provide one stop information system for public and government who should protect sea from oil spill accident. The development of multi user based information system permits integrated handling of real time meteorological data from external ftp. A server-client GIS based model is integrated on the basis of real time database and ESI map to provide the result of the oil spill prediction model. End users can access through the client interface and request analysis such as oil spill prediction and GIS functions on the network as their own purpose.

• Structural Engineering and Mechanics
• /
• 제9권6호
• /
• pp.557-568
• /
• 2000

As time-variant reliability approaches become increasingly used for service life prediction of the aging infrastructure, the demand for computer solution methods continues to increase. Effcient computer techniques have become well established for the reliability analysis of structural systems. Thus far, however, this is largely limited to time-invariant reliability problems. Therefore, the requirements for time-variant reliability prediction of deteriorating structural systems under time-variant loads have remained incomplete. This study presents a computer program for $\underline{REL}$iability of $\underline{T}$ime-Variant $\underline{SYS}$tems, RELTSYS. This program uses a combined technique of adaptive importance sampling, numerical integration, and fault tree analysis to compute time-variant reliabilities of individual components and systems. Time-invariant quantities are generated using Monte Carlo simulation, whereas time-variant quantities are evaluated using numerical integration. Load distribution and post-failure redistribution are considered using fault tree analysis. The strengths and limitations of RELTSYS are presented via a numerical example.

• 로봇학회논문지
• /
• 제11권4호
• /
• pp.277-284
• /
• 2016

A new prediction scheme has been proposed for the robust teleoperation in a non-visible environment. The positioning error caused by the time delay in the non-visible environment has been compensated for by the Smith predictor and the sensory data have been estimated by the Grey model. The Smith predictor is effective for the compensation of the positioning error caused by the time delay with a precise system model. Therefore the dynamic model of a mobile robot has been used in this research. To minimize the unstable and erroneous states caused by the time delay, the estimated sensor data have been sent to the operator. Through simulations, the possibility of compensating the errors caused by the time delay has been verified using the Smith predictor. Also the estimation reliability of the measurement data has been demonstrated. Robust teleoperations in a non-visible environment have been performed with a mobile robot to avoid the obstacles effective to go to the target position by the proposed prediction scheme which combines the Smith predictor and the Grey model. Even though the human operator is involved in the teleoperation loop, the compensation effects have been clearly demonstrated.

• 전기학회논문지
• /
• 제65권12호
• /
• pp.2084-2093
• /
• 2016

It is crucial to predict water temperature for aquatic ecosystem studies and management. In this paper, we first address challenging issues in predicting water temperature in a real time manner and propose a distributed computing model to address such issues. Then, we present an Artificial Neural Network (ANN)-based water temperature prediction model developed for the Soyang River and a cyberinfrastructure system called WT-Agabus to run such prediction models in an automated and real time manner. The ANN model is designed to use only weather forecast data (air temperature and rainfall) that can be obtained by invoking the weather forecasting system at Korea Meteorological Administration (KMA) and therefore can facilitate the automated and real time water temperature prediction. This paper also demonstrates how easily and efficiently the real time prediction can be implemented with the WT-Agabus prototype system.

• 한국정보과학회:학술대회논문집
• /
• 한국정보과학회 1998년도 가을 학술발표논문집 Vol.25 No.2 (2)
• /
• pp.360-362
• /
• 1998

In the original mixtures of experts framework, the parameters of the network are determined by gradient descent, which is naturally slow. In [2], the Expectation-Maximization(EM) algorithm is used instead, to obtain the network parameters, resulting in substantially reduced training times. This paper presents the new EM algorithm for prediction. We show that an Efficient training algorithm may be derived for the HME network. To verify the utility of the algorithm we look at specific examples in time series prediction. The application of the new EM algorithm to time series prediction has been quiet successful.

• 한국항행학회논문지
• /
• 제19권3호
• /
• pp.199-206
• /
• 2015

IGS에서는 실시간 정밀 측위에 사용할 수 있도록 궤도 예측값인 IGU (IGS Ultra-rapid)와 실시간 궤도 추정값인 RTS (real-time service) 보정정보를 제공한다. IGU는 데이터 지연시간이 없지만, RTS는 5~30 초의 지연시간을 갖기 때문에 실시간으로 사용하기 위해선 지연시간만큼 예측이 필요하다. 본 논문에서는 실시간 사용 측면에서 RTS와 IGU의 성능 분석을 수행하였다. 한반도 내에서 RTS 제공 비율을 파악하기 위하여 한반도에서 관측되는 위성 대비 RTS 제공 비율을 계산하였으며 그 결과 99.3%로 나타났다. RTS의 정확도를 확인하기 위해 보정정보를 방송궤도력에 적용하여 오차를 분석하였으며 이 때 3D 궤도 RMS 오차는 0.043 m으로 나타났다. 실시간 사용 측면에서 RTS와 IGU를 비교하였는데, 실시간으로 가정하였기 때문에 IGU는 예측 정보만 이용하였고, RTS는 데이터 지연시간동안 다항식 모델로 예측을 수행하였다. RTS와 IGU를 1초 간격으로 각각 외삽, 보간을 수행하였고 그 결과 궤도 예측 성능은 비슷하였으며 시계 예측 성능은 RTS가 0.13 m 더 뛰어났다.