• Communications for Statistical Applications and Methods
• /
• v.3 no.2
• /
• pp.227-237
• /
• 1996

비중심 ${\chi}^2$분포의 누적분포함수의 계산은 ${\chi}^2$검정에서 요구되고 있는 새로운 접근방법으로 신경망 이론을 적용하기 위하여 입력층의 입력노드가 세개, 출력증의 축력노드가 한개 그리고 한개의 은닉층으로 구성된 다층 퍼셉트론 네트워크부터 역전파 알고리즘을 개발하여 비중심${\chi}^2$분포의 확률계산을 시도하였다. 정확성과 계산속도를 고려하여 기존의 방법과 비교한 결과 효율적임을 알 수 있다.

• Journal of Digital Contents Society
• /
• v.18 no.1
• /
• pp.183-190
• /
• 2017

This paper proposes an audio event detection method using Deep Neural Networks (DNN). The proposed method applies Feed Forward Neural Network (FFNN) to generate output probabilities of twenty audio events for each frame. Mel scale filter bank (FBANK) features are extracted from each frame, and its five consecutive frames are combined as one vector which is the input feature of the FFNN. The output layer of FFNN produces audio event probabilities for each input feature vector. More than five consecutive frames of which event probability exceeds threshold are detected as an audio event. An audio event continues until the event is detected within one second. The proposed method achieves as 71.8% accuracy for 20 classes of the UrbanSound8K and the BBC Sound FX dataset.

• Journal of Advanced Navigation Technology
• /
• v.13 no.2
• /
• pp.208-213
• /
• 2009

This paper presents an adaptive call admission control(CAC) algorithm based on a target handoff call dropping probability in mobile wireless environments. This method uses a neural network for predicting and reserving the bandwidth demands for handoff calls and new calls. The amount of bandwidth to be reserved is adaptively adjusted by a target value of handoff call dropping probability(CDP). That is, if the handoff CDP exceeds the a target CDP value, the bandwidth to be reserved should be increased to reduce the handoff dropping probability below a target value. The proposed method is intended to prevent from increasing handoff call dropping probability when bandwidth to be reserved is not enough for handoff calls due to an uncertain prediction. Our simulations compare the handoff CDP in proposed CAC with that of an existing CAC. Results show that the proposed method sustains handoff call dropping probability below our target value.

• Proceedings of the Earthquake Engineering Society of Korea Conference
• /
• 2006.03a
• /
• pp.382-389
• /
• 2006

구조 재료와 시공기술의 발달로 구조물은 높고 길게 설계할 수 있게 되었으나, 그에 따른 진동 문제와 사용성에 관한 문제가 발생하였고 구조물의 과다한 변위는 구조물에 심각한 손상을 발생 시켰다. 이러한 구조물의 진동 문제를 해결하기 위하여 본 논문에서는 확률신경망이론을 사용한 구조물의 능동제어방법을 제안하였다. 구조물의 제어를 위하여 LQR 제어알고리즘을 이용하여 구조물의 상태벡터와 제어력을 구한 후, 상태벡터를 입력으로 제어력을 출력으로 하는 확률신경망의 훈련패턴을 구성하였다. 제안된 방법을 사용하여 지진하중을 받는 3층 빌딩구조물을 제어하였고, 기존의 인공신경망의 제어 결과와 비교하였다.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.24 no.12B
• /
• pp.2447-2454
• /
• 1999

This paper proposes a N-APPEM(Nonlinear A Posteriori Probability Estimation Method) with a frame normalization method to conventional classification network to increase speech recognition ability. It also tests the recognition ability of the classification and prediction neural networks for the Korean isolated digits. From the experimental results, the prediction network with MLP(Multi-Layer Perceptron) achieves the highest recognition ability of 98.0%. The prediction requires very complicated networks increased linearly with the number of incoming speech categories. However, the classification network with the N-APPEM and the normalization improves the recognition ability up to 85.5% with a sin81e network, which is almost 12.0% improvement.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.36S no.1
• /
• pp.123-132
• /
• 1999

In this paper, a quadratic sigmoid neural equalizer(QSNE) is proposed to improve the performance of conventional neural equalizer in terms of bit error probability by using a quadratic sigmoid function as the activation function of neural networks. Conventional neural equalizers which have been used to compensate for nonlinear distortions adopt the sigmoid function. In the case of sigmoid neural equalizer, each neuron has one linear decision boundary. So many neurons are required when the neural equalizer has to separate complicated structure. But in case of the proposed QSNF and quadratic sigmoid neural decision feedback equalizer(QSNDFE), each neuron separates decision region with two parallel lines. Therefore, QSNE and QSNDFE have better performance and simpler structure than the conventional neural equalizers in terms of bit error probability. When the proposed QSNDFE is applied to communication systems and digital magnetic recording systems, it is an improvement of approximately 1.5dB~8.3dB in signal to moise ratio(SNR) over the conventional decision feedback equalizer(DEF) and neural decision feedback equalizer(NDFE). As intersymbol interference(ISI) and nonlinear distortions become severer, QSNDFE shows astounding SNR shows astounding SNR performance gain over the conventional equalizers in the same bit error probability.

• Journal of the Korean Data and Information Science Society
• /
• v.12 no.2
• /
• pp.11-25
• /
• 2001

There are several algorithms for classification in modeling relations, patterns, and rules which exist in data. We learn to classify objects on the basis of instances presented to us, not by being given a set of classification rules. The Bayesian learning uses the probability distribution to express our knowledge about unknown parameters and update our knowledge by the law of probability as the evidence gathered from data. Also, the neural network models are designed for predicting an unknown category or quantity on the basis of known attributes by training. In this paper, we compare the misclassification error rates of Bayesian Neural Network method with those of other classification algorithms, CHAID, CART, and QUBST using several data sets.

• The Journal of the Korea Contents Association
• /
• v.7 no.1
• /
• pp.65-72
• /
• 2007

In this study, we used nondestructive test based on ultrasonic test as inspection method and compared backpropagation neural network(BPNN) with probabilistic neural network(PNN) as pattern recognition algorithm of weld flaws. For this purpose, variables are applied the same to two algorithms. Where, feature variables are zooming flaw signals of reflected whole signals from weld flaws in time domain. Through this process, we compared advantages/ disadvantages of two algorithms and confirmed application methods of two algorithms.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2021.06a
• /
• pp.156-156
• /
• 2021

본 연구는 지역빈도해석을 기반으로한 인공신경망 모델과 기존에 널리 사용되는 방법인 홍수지수법의 성능을 몬테카를로 시뮬레이션을 이용하여 평가하였다. 컴퓨터 기술이 발달함에 따라 인공지능에 대한 접근성이 좋아지며 수문학을 포함한 다양한 분야에 적용되고 있다. 인공지능을 이용하여 강수량 및 유량 등 다양한 수문자료에 대한 예측이 이루어지고 있으나 빈도해석에 관한 연구는 비교적 적다. 본 연구에서 사용된 인공 지능 모델은 대상 지점의 지형학적 자료와 수문학적 자료를 이용하여 인공신경망을 통해 지점의 확률강우량(QRT-ANN) 및 확률분포형의 매개변수 (PRT-ANN)를 추정한다. 지형학적 자료로는 위도, 경도 그리고 고도가 사용되었으며 수문학적 자료로는 대상 지점의 최근 30년 일일연최대강우량을 사용하였다. 지역빈도해석의 정확도는 지역 내 통계적 특성이 비슷한 지점들이 포함되면 될수록 높아진다. 통계적 특성으로는 불일치 척도, 이질성 척도, 적합성 척도가 있으며 다양한 조건의 통계적 특성에 따른 세 개의 지역빈도해석 방법의 성능을 평가하고자 하였다. 대상 지역 내 n개의 지점이 있다고 가정하였을 때, 홍수지수법의 경우 n-1개의 지점으로 추정한 지역 성장곡선을 이용하여 나머지 1개 지점의 확률강우량을 산정할 수 있으며 인공신경망 모델들 또한 n-1개 지점들의 자료를 이용하여 모델을 구축한 뒤 나머지 지점의 확률강우량 및 확률분포형의 매개변수를 예측할 수 있다. PRT-ANN의 경우 예측된 매개변수를 이용하여 확률강우량을 산정하며 시뮬레이션 시행마다 발생시킨 자료의 지점빈도해석 결과에 대한 나머지 세 방법의 평균 제곱근 상대오차 (Relative root mean square error, RRMSE)를 계산하였다. 몬테카를로 시뮬레이션을 이용한 성능 분석을 통하여 관측값의 다양한 통계적 특성에 맞는 지역빈도해석 방법을 제시할 수 있을 것으로 판단된다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.25 no.4
• /
• pp.501-507
• /
• 2021

Neural networks with differentiable activation functions are differentiable with respect to input variables. We improve the approximation capability of neural networks by using the gradient and Hessian of neural networks to satisfy the differential equations of the problems of interest. We apply differential neural networks to the pricing of financial options, where stochastic differential equations and the Black-Scholes partial differential equation represent the differential relation of price of option and underlying assets, and the first and second derivatives of option price play an important role in financial engineering. The proposed neural network learns - (a) the sample paths of option prices generated by stochastic differential equations and (b) the Black-Scholes equation at each time and asset price. Experimental results show that the proposed method gives accurate option values and the first and second derivatives.