• ETRI Journal
• /
• 제32권5호
• /
• pp.795-800
• /
• 2010

Embedding a large vocabulary speech recognition system in mobile devices requires a reduced acoustic model obtained by eliminating redundant model parameters. In conventional optimization methods based on the minimum description length (MDL) criterion, a binary Gaussian tree is built at each state of a hidden Markov model by iteratively finding and merging similar mixture components. An optimal subset of the tree nodes is then selected to generate a downsized acoustic model. To obtain a better binary Gaussian tree by improving the process of finding the most similar Gaussian components, this paper proposes a new distance measure that exploits the difference in likelihood values for cases before and after two components are combined. The mixture weight of Gaussian components is also introduced in the component merging step. Experimental results show that the proposed method outperforms MDL-based optimization using either a Kullback-Leibler (KL) divergence or weighted KL divergence measure. The proposed method could also reduce the acoustic model size by 50% with less than a 1.5% increase in error rate compared to a baseline system.

• 로봇학회논문지
• /
• 제5권1호
• /
• pp.70-76
• /
• 2010

This paper addresses a design issue of "model complexity and performance trade-off" in the application of bandwidth extension (BWE) methods to the intra-frame predictivevector quantization problem of wideband speech. It discusses model-based linear and non-linear prediction methods and presents a comparative study of them in terms of prediction gain. Through experimentation, the general trend of saturation in performance (with the increase in model complexity) is observed. However, specifically, it is also observed that there is no significant difference between HMM and GMM-based BWE functions.

• 음성과학
• /
• 제13권4호
• /
• pp.177-186
• /
• 2006

In an effort to enhance the quality of feature vector classification and thereby reduce the recognition error rate of the speaker-independent speech recognition, we employ the Mahalanobis distance in the calculation of the similarity measure between feature vectors. It is assumed that the metric matrix of the Mahalanobis distance be diagonal for the sake of cost reduction in memory and time of calculation. We propose that the diagonal elements be given in terms of the variations of the feature vector components. Geometrically, this prescription tends to redistribute the set of data in the shape of a hypersphere in the feature vector space. The idea is applied to the speech recognition by hidden Markov model with fuzzy vector quantization. The result shows that the recognition is improved by an appropriate choice of the relevant adjustable parameter. The Viterbi score difference of the two winners in the recognition test shows that the general behavior is in accord with that of the recognition error rate.

• 대한수학회지
• /
• 제59권2호
• /
• pp.311-335
• /
• 2022

This paper treats Merton's classical portfolio optimization problem for a market participant who invests in safe assets and risky assets to maximize the expected utility. When the state process is a d-dimensional Markov diffusion, this problem is transformed into a problem of solving a Hamilton-Jacobi-Bellman (HJB) equation. The main purpose of this paper is to solve this HJB equation by a deep learning algorithm: the deep Galerkin method, first suggested by J. Sirignano and K. Spiliopoulos. We then apply the algorithm to get the solution to the HJB equation and compare with the result from the finite difference method.

• Current Research on Agriculture and Life Sciences
• /
• 제7권
• /
• pp.55-66
• /
• 1989

낙동강수계중(洛東江水系中)의 3개(個) 중대(中大) 유역(流域)인 안동댐 유역(流域)($1612.8km^2$), 임하유역($1964.8km^2$), 선산($979.3km^2$)을 대상(對象)으로 Horton과 Strahler의 하천차수법칙(河川次數法則)에 의해 하천(河川)을 분류(分類)하고, 1 : 50,000 지형도(地形圖)를 이용(利用)하여 각(各) 하천차수별(河川次數別) 지상인자(地相因子)를 구(求)하여 이들 지상인자(地相因子)들에 의해 Markov확률과정(確率過程)을 도입(導入)해서 구(求)한 순간단위도(瞬間單位圖)를 이용(利用)한 방법(方法)과 SCS무차원단위도법(無次元單位圖法)에 의해 홍수량(洪水量)을 구(求)한 결과(結果)를 요약하면 다음과 같다. 1. 대상유역(對象流域)에서 Horton의 하천분류법(河川分流法)에 따른 지상인자(地相因子)들의 유역특성(流域特性)인 분기비(分岐比), 길이비(比), 면적비(面積比)는 각각 3.6~5.5, 1.8~3.8, 5.1~6.4의 범위(範圍)로써 Horton이 제안(提案)한 값과 거의 일치(一致)하였다. 2. Markov 확률과정(確率科程)을 이용(利用)한 순간단위도(瞬間單位圖)에서 홍수량(洪水量)은 0~7%, 홍수도달시간(洪水到達時間)은 0~2시간(時間)의 오차로 양호한 값이 나타난데 반하여, SCS무차원단위법(無次元單位法)으로 구(求)한 홍수량(洪水量)은 10~40%, 홍수도달시간(洪水到達時間)은 0~4시간(時間)의 비교적 큰 오차가 나타났다.

• 지능정보연구
• /
• 제17권3호
• /
• pp.187-201
• /
• 2011

방향성(Direction)과 변동성(Volatility)에 대한 분석은 증권투자를 위한 시장분석의 기초가 된다. 변동성분석이 옵션 투자에서 중요하다면 주식이나 주가지수선물투자는 방향성분석에 의하여 투자성과가 결정된다. 기존의 금융분석에서 기계학습을 이용한 방향성에 대한 연구는 주가나 투자위험의 예측을 중심으로 이루어졌으며, 최근에 와서야 실전투자를 위한 매매시스템(trading system) 개발에 대한 연구가 이루어지고 있다. 인공지능형 주가예측모형에서는 ANN(artificial neural networks), fuzzy system, SVM(Support Vector Machine) 등의 기법이 주로 활용되고 있다. 본 연구에서는 방향성매매를 위한 지능형 기계학습방법 중에서도 패턴인식에서 좋은 성과를 보이고 있는 은닉마코프 모형(Hidden Markov Model)을 이용한다. 실무적으로는 방향성 예측을 위해 주로 주가의 추세분석(Trend Analysis)을 활용한다. 다양한 기술적 지표를 이용한 추세분석에 기반한 시스템트레이딩(System Trading) 기법은 실전투자에서 점차 확대추세에 있다. 본 연구에서는 시스템트레이딩 기법 중 실무에서 많이 이용되는 이동평균교차전략(moving average cross)에 연속 은닉마코프모형을 적용한 지능형 매매시스템을 제안하고, 실제 주가자료를 이용한 시뮬레이션 결과를 제시한다. 세계적 선물시장으로 성장한 KOSPI200 선물시장에서 제안된 매매시스템의 장기간의 투자성과를 분석하기 위하여 지난 21년 동안의 KOSPI200 주가지수자료를 실증 분석하였다. 분석결과는 KOSPI200 주가지수선물의 방향성매매에서 제안된 CHMM기반 지능형 매매시스템이 실전에서 일반적으로 활용되는 시스템트레이딩 기법의 투자성과를 개선할 수 있음을 보여주었다.

• Asian-Australasian Journal of Animal Sciences
• /
• 제22권1호
• /
• pp.124-130
• /
• 2009

Estimating genetic interaction effects in animal genomics would be one of the most challenging studies because the phenotypic variation for economically important traits might be largely explained by interaction effects among multiple nucleotide sequence variants under various environmental exposures. Genetic improvement of economic animals would be expected by understanding multi-locus genetic interaction effects associated with economic traits. Most analyses in animal breeding and genetics, however, have excluded the possibility of genetic interaction effects in their analytical models. This review discusses a historical estimation of the genetic interaction and difficulties in analyzing the interaction effects. Furthermore, two recently developed methods for assessing genetic interactions are introduced to animal genomics. One is the restricted partition method, as a nonparametric grouping-based approach, that iteratively utilizes grouping of genotypes with the smallest difference into a new group, and the other is the Bayesian method that draws inferences about the genetic interaction effects based on their marginal posterior distributions and attains the marginalization of the joint posterior distribution through Gibbs sampling as a Markov chain Monte Carlo. Further developing appropriate and efficient methods for assessing genetic interactions would be urgent to achieve accurate understanding of genetic architecture for complex traits of economic animals.

• 한국멀티미디어학회논문지
• /
• 제20권4호
• /
• pp.593-605
• /
• 2017

In applications which are covered wide range, it is possible that one or more number of Unmanned Aerial Vehicle(UAV) squadrons are used to perform a mission. In this case, it is most important to communicate seamlessly between the UAV squadrons. In this paper, we applied the modified OLSR(OSLR-Pds) which can prediction for state of the link for the communication in UAV squadron, and applied the modified AOMDV which can build multi-path for the communication between UAV Squadrons. The mobility of nodes are modeled using Gauss-Markov algorithm, and relative speed between nodes were calculated by derive equation of movement, and thereby we can predict link state for in a squadron and between squadrons. An experiment for comparing AODV, AOMDV and the proposed routing protocol was conducted by three factors such as packet delivery ratio, end to end delay, and routing overhead. In experiment result, we make sure that the proposed protocol performance are superior in these three factors. However, if the density of the nodes constituting FANET are too low, and if the moving speed of node is very slow, there is no difference to others protocols.

• 한국정보과학회:학술대회논문집
• /
• 한국정보과학회 1998년도 가을 학술발표논문집 Vol.25 No.2 (2)
• /
• pp.384-386
• /
• 1998

The use of hand gesture provides an attractive alternative to cumbersome interface devices for human-computer interaction(HCI). Many methods hand gesture recognition using visual analysis have been proposed such as syntactical analysis, neural network(NN), Hidden Markov Model(HMM) and so on. In our research, a HMMs is proposed for alphabetical hand gesture recognition. In the preprocessing stage, the proposed approach consists of three different procedures for hand localization, hand tracking and gesture spotting. The hand location procedure detects the candidated regions on the basis of skin-color and motion in an image by using a color histogram matching and time-varying edge difference techniques. The hand tracking algorithm finds the centroid of a moving hand region, connect those centroids, and thus, produces a trajectory. The spotting a feature database, the proposed approach use the mesh feature code for codebook of HMM. In our experiments, 1300 alphabetical and 1300 untrained gestures are used for training and testing, respectively. Those experimental results demonstrate that the proposed approach yields a higher and satisfying recognition rate for the images with different sizes, shapes and skew angles.

• 한국측량학회:학술대회논문집
• /
• 한국측량학회 2010년 춘계학술발표회 논문집
• /
• pp.429-431
• /
• 2010

The hybrid geoid model should be determined by fitting the gravimetric geoid to the geometric geoid which were presented the local vertical level. Therefore, it is necessary to find firstly the optimal scheme for improving the accuracy of gravimetric geoid in order to development the high-precision hybrid geoid model. Through finding the optimal scheme for determining the each part of gravimetric geoid, the most accurate gravimetric geoid model in Korea will be developed when the EIGEN-CG03C model to degree 360, 4-band spherical FFT and RTM reduction methods were used for determining the long, middle and short-frequency part of gravimetric geoid respectively. Finally, we developed the hybrid geoid model around Korea by correcting to gravimetric geoid with the correction term. The correction term is modelled using the difference between GPS/Levelling derived geoidal heights and gravimetric geoidal heights. The stochastic model used in the calculation of correction term is the LSC technique based on second-order Markov covariance function. 503 GPS/Levelling data were used to model the correction term. The degree of LSC fitting to the final hybrid geoid model in Korea was evaluated as 0.001m ${\pm}0.054m$.