• 한국산업정보학회논문지
• /
• 제15권2호
• /
• pp.49-57
• /
• 2010

본 논문에서는 역전파 신경망 알고리즘(BPNN: Back Propagation Neural Network)과 Singular Value Decomposition(SVD)를 이용하는 한글 문서 분류 시스템을 제안한다. BPNN은 학습을 통하여 만들어진 네트워크를 이용하여 문서분류를 수행한다. 이 방법의 어려움은 분류기에 입력되는 특징 공간이 너무 크다는 것이다. SVD를 이용하면 고차원의 벡터를 저차원으로 줄일 수 있고, 또한 의미있는 벡터 공간을 만들어 단어 사이의 중요한 관계성을 구축할 수 있다. 본 논문에서 제안한 BPNN의 성능 평가를 위하여 한국일보-2000/한국일보-40075 문서범주화 실험문서집합의 데이터 셋을 이용하였다. 실험결과를 통하여 BPNN과 SVD를 사용한 시스템이 한글 문서 분류에 탁월한 성능을 가지는 것을 보여준다.

• ETRI Journal
• /
• 제30권4호
• /
• pp.546-554
• /
• 2008

Phase-locked loops (PLLs) are among the most important mixed-signal building blocks of modern communication and control circuits, where they are used for frequency and phase synchronization, modulation, and demodulation as well as frequency synthesis. The growing popularity of PLLs has increased the need to test these devices during prototyping and production. The problem of distinguishing and classifying the responses of analog integrated circuits containing catastrophic faults has aroused recent interest. This is because most analog and mixed signal circuits are tested by their functionality, which is both time consuming and expensive. The problem is made more difficult when parametric variations are taken into account. Hence, statistical methods and techniques can be employed to automate fault classification. As a possible solution, we use the back propagation neural network (BPNN) to classify the faults in the designed charge-pump PLL. In order to classify the faults, the BPNN was trained with various training algorithms and their performance for the test structure was analyzed. The proposed method of fault classification gave fault coverage of 99.58%.

• 대한전기학회논문지:시스템및제어부문D
• /
• 제51권7호
• /
• pp.327-333
• /
• 2002

In this study, we developed a snoring detection algorithm that detects snores automatically. It consists of preprocessing and snoring detection part. The preprocessing part is composed of a noise removal part using spectrum subtraction, and segmentation part, and computation part of temporal and spectral features. And the snoring detection part decides whether detected blocks are snores with BPNN(Back-Propagation Neural Network). BPNN with one hidden layer and one output layer, is trained with data of 7 subjects and tested with data of 11 subjects of total 18 subjects. The proposed algorithm showed a Sensitivity of 90.41% and a Predictive Positive Value of 84.95%.

• 한국지능시스템학회논문지
• /
• 제8권5호
• /
• pp.39-44
• /
• 1998

이동 로봇의 운행을 위해서 이동 로봇의 절대 위치를 결정하는 것이 중요하다. 본 논문에서는 신경회로망을 이용하여 랜드마크의 영상을 통해 이동 로봇의 위치를 결정하는 방법을 제안한다. 픽셀의 불확시한 값, 부정확한 카메라 조정과 렌즈의 왜곡으로 인해 이동 로봇의 위치를 결정에 있어서 위치 오차가 생기게 된다. 이러한 오차를 줄이기 위해서 BPNN(Back Propagation Neural Network)를 사용하는 방법을 제안한다. 기존의 방법과 비교하여 우수성을 보여주기 위해서 실험결과를 보여준다.

• 소성∙가공
• /
• 제30권3호
• /
• pp.125-133
• /
• 2021

In this study, response surface method (RSM), back propagation neural network (BPNN), and genetic algorithm (GA) were used for modeling and multi-objective optimization of the parameters of AA5052-H32 in incremental sheet forming (ISF). The goal of optimization is to determine the maximum forming angle and minimum surface roughness, while varying the production process parameters, such as tool diameter, tool spindle speed, step depth, and tool feed rate. A Box-Behnken experimental design (BBD) was used to develop an RSM model and BPNN model to model the variations in the forming angle and surface roughness based on variations in process parameters. Subsequently, the RSM model was used as the fitness function for multi-objective optimization of the ISF process the GA. The results showed that RSM and BPNN can be effectively used to control the forming angle and surface roughness. The optimized Pareto front produced by the GA can be utilized as a rational design guide for practical applications of AA5052 in the ISF process

• 대한토목학회논문집
• /
• 제26권2C호
• /
• pp.81-88
• /
• 2006

연약지반의 설계정수로 사용되는 비배수전단강도 및 선행압밀하중의 예측을 위해 전국적으로 산재해 있는 6개의 연약지반 대상구역의 실험결과를 이용하여 역전파학습알고리즘을 통해 학습 및 예측을 실시하였다. 실험결과치와 신경망학습의 결과치는 상관계수 0.9이상의 값을 나타냄으로서 높은 상관성를 나타내었으며 자연함수비, 간극비, 비중, 세립토의 함유율은 상관성을 높이는데 상당한 기여를 하는 것으로 나타났다. 본 연구를 통해 연약지반개량공법설계시 충분한 양질의 자료만 확보할 수 있다면 다양한 지반의 물성치를 인공신경망을 통해 효율적으로 예측할 수 있다는 것을 확인하였다.

• International Journal of Quality Innovation
• /
• 제7권3호
• /
• pp.58-69
• /
• 2006

This paper presents neural network-based recognition system for automatic light emitting diode (LED) inspection. The back-propagation neural network (BPNN) is proposed and tested. The current-voltage (I-V) characteristic data of LED from the inspection process is used for the network training and testing. This study selects 300 random samples as network training and employs 100 samples as network testing. The experimental results show that if the classification work is done well, the accuracy of recognition is 100%, and the testing speed of the proposed recognition system is almost one half faster than the traditional inspection system does. The proposed neural-network approach is successfully demonstrated by real data sets and can be effectively developed as a recognition system for a practical application purpose.

• Wind and Structures
• /
• 제30권3호
• /
• pp.289-298
• /
• 2020

Gusts generated by downburst have caused a great variety of structural damages in many regions around the world. It is of great significance to accurately evaluate the downburst-induced wind load on high-rise building for the wind resistance design. The main objective of this paper is to propose a computational modeling approach which can satisfactorily predict the mean and fluctuating wind pressure coefficients induced by downburst on high-rise building surfaces. In this study, using an impinging jet to simulate downburst-like wind, and simultaneous pressure measurements are obtained on a high-rise building model at different radial locations. The model test data are used as the database for developing back propagation neural network (BPNN) models. Comparisons between the BPNN prediction results and those from impinging jet test demonstrate that the BPNN-based method can satisfactorily and efficiently predict the downburst-induced wind pressure coefficients on single and overall surfaces of high-rise building at various radial locations.

• International Journal of Naval Architecture and Ocean Engineering
• /
• 제12권1호
• /
• pp.354-366
• /
• 2020

This work mainly focuses on determination of the fully plastic J-integral solutions for welded center cracked plates subjected to remote tension loading. Detailed three-dimensional elasticeplastic Finite Element Analyses (FEA) were implemented to compute the fully plastic J-integral along the crack front for a wide range of crack geometries, material properties and weld strength mismatch ratios for 900 cases. According to the database generated from FEA, Back-propagation Neural Network (BPNN) model was proposed to predict the values and distributions of fully plastic J-integral along crack front based on the variables used in FEA. The determination coefficient R2 is greater than 0.99, indicating the robustness and goodness of fit of the developed BPNN model. The network model can accurately and efficiently predict the elastic-plastic J-integral for weld centerline crack, which can be used to perform fracture analyses and safety assessment for welded center cracked plates with varying strength mismatch conditions under uniaxial loading.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2000년도 제15차 학술회의논문집
• /
• pp.20-20
• /
• 2000

Back-propagation neural network (BPNN) is the most prevalently used paradigm in modeling semiconductor manufacturing processes, which as a neuron activation function typically employs a bipolar or unipolar sigmoid function in either hidden and output layers. In this study, applicability of another linear function as a neuron activation function is investigated. The linear function was operated in combination with other sigmoid functions. Comparison revealed that a particular combination, the bipolar sigmoid function in hidden layer and the linear function in output layer, is found to be the best combination that yields the highest prediction accuracy. For BPNN with this combination, predictive performance once again optimized by incrementally adjusting the gradients respective to each function. A total of 121 combinations of gradients were examined and out of them one optimal set was determined. Predictive performance of the corresponding model were compared to non-optimized, revealing that optimized models are more accurate over non-optimized counterparts by an improvement of more than 30%. This demonstrates that the proposed gradient-optimized teaming for BPNN with a linear function in output layer is an effective means to construct plasma models. The plasma modeled is a hemispherical inductively coupled plasma, which was characterized by a 24 full factorial design. To validate models, another eight experiments were conducted. process variables that were varied in the design include source polver, pressure, position of chuck holder and chroline flow rate. Plasma attributes measured using Langmuir probe are electron density, electron temperature, and plasma potential.