• 대한용접접합학회:학술대회논문집
• /
• 대한용접접합학회 2000년도 특별강연 및 춘계학술발표대회 개요집
• /
• pp.251-255
• /
• 2000

Recently, the improvement of computer capacities and artificial intelligence ware caused to employ for prediction of residual stresses and strength evaluation. There are a lot of researches regarding the measurement and prediction of residual stresses for weldment using a neural network in the advanced countries, but in our country, a neural network as a technical part, has only been used on the possibilities of employment for welding area. Furthermore, the relationship between residual stress and process parameters using a neural network was wholly lacking. Therefore development of a new technical method for the optimized process parameters on the reduction of residual stress and applyment of real-time production line should be developed. The objectives of this paper is to measure the residual stress of butt welded specimen using strain gage sectioning method and to apply them to a neural network for prediction of residual stresses on a given process parameter. Also, the assessment of the developed system using a neural network was carried out

• 대한전자공학회논문지SP
• /
• 제40권1호
• /
• pp.139-147
• /
• 2003

동적 신경망은 신호예측과 같이 temporal 신호처리가 요구되는 여러 분야에 적용되어 왔다. 본 논문에서는 다층 리커런트 신경망(RNN)의 동특성을 향상시키기 위해 지역 궤환 신경망(LRNN)과 광역 궤환 신경망(CRNN)으로 구성된 합성 신경망을 제안하고, 적응필터로 제안된 신경망을 사용하여 비선형 적응예측을 다루고 있다. 합성 신경망은 LRNN으로 IIR-MLP와 CRNN으로 Elman RNN 신경망으로 구성되어 있다. 제안된 신경망은 비선형 신호예측을 통해 평가되었으며, 예측 성능의 상대적인 비교를 위해 Elman RNN과 IIR-MLP 신경망과 상호 비교하였다. 실험결과에 의하면 합성 신경망은 수렴속도과 정확도에서 더 우수한 성능을 보여줌으로써, 제안된 신경망이 기존의 다층 리커런트 신경망보다 비정적 신호에 대한 비선형 예측에 더 효과적인 예측모델임을 확인하였다.

• 한국항행학회논문지
• /
• 제25권5호
• /
• pp.409-414
• /
• 2021

미세먼지에 대한 심각성이 사회적으로 대두됨에 따라 대중들은 미세먼지 예보에 대한 정보의 높은 신뢰성을 요구하고 있다. 이에 따라 다양한 신경망 알고리즘을 이용하여 미세먼지 예측을 위한 연구가 활발히 진행되고 있다. 본 논문에서는 미세먼지 예측을 위해 다양한 알고리즘으로 연구되고 있는 신경망 알고리즘들 중 대표적인 알고리즘들의 예측 성능 비교를 진행하였다. 신경망 알고리즘 중 DNN(deep neural network), RNN(recurrent neural network), LSTM(long short-term memory)을 이용하였으며, 하이퍼 파라미터 탐색을 이용하여 최적의 예측 모델을 설계하였다. 각 모델의 예측 성능 비교 분석 결과, 실제 값과 예측 값의 변화 추이는 전반적으로 좋은 성능을 보였다. RMSE와 정확도를 기준으로 한 분석에서는 DNN 예측 모델이 다른 예측 모델에 비해 예측 오차에 대한 안정성을 갖는 것을 확인하였다.

• 융합신호처리학회논문지
• /
• 제3권4호
• /
• pp.62-66
• /
• 2002

SOCMAC 신경망에 의하여 Mackey-Glass의 비선형 시계열 예측을 시도하였다 다차원 연속 입력 변수를 가지는 문제는 요구되는 기억용량의 규모가 너무 커서 CMAC에서는 일반적으로 취급이 곤난한 대상이었으나 SOCMAC에서는 이것이 가능함을 보였다. 또한 학습과정에서 수용영역(receptive field)을 가변으로 하는 개선된 방법을 제시하였다. 예측오차는 TDNN(time-delayed neural network)이나 BP(back-propagation) 수준이었다.

• 한국음향학회지
• /
• 제20권8호
• /
• pp.12-18
• /
• 2001

본문에서는 예측형 회귀신경망과 HMM (Hidden Markov Model)의 하이브리드 네트워크인 회귀신경망 예측 HMM을 구성하였다. 회귀신경망 예측 HMM은 예측형 회귀신경망을 HMM의 각 상태마다 예측기로 정의하여 일정치인 평균벡터 대신에 과거의 특징벡터의 영향을 받아 동적으로 변화하는 신경 망에 의한 예측치를 이용하므로 학습패턴 설정자체가 시변성을 반영하는 동적 네트워크의 특성을 가진다. 따라서 음성과 같은 시계열 패턴의 인식에 유리하다. 회귀신경망 예측 HMM은 예측형 회귀신경망의 구조에 따라 Elman망 예측 HMM과 Jordan망 예측 HMM으로 구분하였다. 실험에서는 회귀신경망 예측 HMM의 상태수를 4, 5, 6으로 증가시켜 각 상태 수별로 예측차수 및 중간층 유니트 수의 변화에 따른 인식성능을 조사하였다. 실험결과 평가용 데이터에 대하여 Elman망 예측 HMM은 상태수가 6이고, 예측차수가 3차, 중간층 유니트의 수가 15차원일 때, Jordan망 예측 HMM의 경우 상태수가 5이고, 예측차수가 3차, 중간층 유니트의 수가 10차원일 때 각각 98.5%로 우수한 결과를 얻었다.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1999년도 하계학술대회 논문집 G
• /
• pp.2863-2865
• /
• 1999

In this paper we present the modeling of ozone prediction system using polynomial neural network. The Polynomial Neural Network is a useful tool for data learning, nonlinear function estimation and prediction of dynamic system. The mechanism of ozone concentration is highly complex, nonlinear, nonstationary. The purposed method shows that the prediction to the ozone concentration based upon a polynomial neural network gives us a good performance for ozone prediction with ability of superior data approximation.

• 한국철도학회:학술대회논문집
• /
• 한국철도학회 2006년도 추계학술대회 논문집
• /
• pp.1405-1410
• /
• 2006

Purpose of this research is that offers basic data for optimized design using neural network method to calculate consolidation settlement in study area. In this research, preformed the neural network method that analyzed the settlement characteristics of soft ground nearby study area. Thus, data base established on ground properties and consolidation settlement of neighboring area. In addition, designed the optimum neural network model for prediction of settlement through network learning and consolidation settlement prediction using consolidation settlement DB and ground properties DB. Optimized neural network model decided by repeated learning for various case of hidden layers. In this study, proposed that the optimized consolidation settlement calculation method using neural network and verified which is the optimized consolidation settlement calculation method using neural network.

• 한국가시화정보학회지
• /
• 제21권1호
• /
• pp.18-25
• /
• 2023

In these days, high dimensional data prediction technology based on neural network shows compelling results in many different kind of field including engineering. Especially, a lot of variants of convolution neural network are widely utilized to develop pixel level prediction model for high dimensional data such as picture, or physical field value from the sensors. In this study, velocity vector field of ideal flow on ship surface is estimated on pixel level by Unet. First, potential flow analysis was conducted for the set of hull form data which are generated by hull form transformation method. Thereafter, four different neural network with a U-shape structure were conFig.d to train velocity vectors at the node position of pre-processed hull form data. As a result, for the test hull forms, it was confirmed that the network with short skip-connection gives the most accurate prediction results of streamlines and velocity magnitude. And the results also have a good agreement with potential flow analysis results. However, in some cases which don't have nothing in common with training data in terms of speed or shape, the network has relatively high error at the region of large curvature.

• 한국멀티미디어학회논문지
• /
• 제18권4호
• /
• pp.524-532
• /
• 2015

All data created in BigData times is included potentially meaning and correlation in data. A variety of data during a day in all society sectors has become created and stored. Research areas in analysis and grasp meaning between data is proceeding briskly. Especially, accuracy of meaning prediction and data imbalance problem between data for analysis is part in course of something important in data analysis field. In this paper, we proposed data prediction model based on data weights and neural network using R for meaning analysis between data. Proposed data prediction model is composed of classification model and analysis model. Classification model is working as weights application of normal distribution and optimum independent variable selection of multiple regression analysis. Analysis model role is increased prediction accuracy of output variable through neural network. Performance evaluation result, we were confirmed superiority of prediction model so that performance of result prediction through primitive data was measured 87.475% by proposed data prediction model.

• 대한전기학회논문지:시스템및제어부문D
• /
• 제51권1호
• /
• pp.39-42
• /
• 2002

For the purpose of modeling EEG signal which has nonstationary and nonlinear dynamic characteristics, this paper propose a state feedback real time recurrent neural network model. The state feedback real time recurrent neural network is structured to have memory structure in the state of hidden layers so that it has arbitrary dynamics and ability to deal with time-varying input through its own temporal operation. For the model test, Mackey-Glass time series is used as a nonlinear dynamic system and the model is applied to the prediction of three types of EEG, alpha wave, beta wave and epileptic EEG. Experimental results show that the performance of the proposed model is better than that of other neural network models which are compared in this paper in some view points of the converging speed in learning stage and normalized mean square error for the test data set.