• Proceedings of the Korean Society of Computer Information Conference
• /
• 2014.07a
• /
• pp.327-328
• /
• 2014

클라우드 컴퓨팅 환경에서는 시스템의 성능 및 비용적인 측면에서 정보 공유의 용이성, 장소의 제약성 최소화, 저장 공간의 효율적 사용을 위해 분산 파일시스템을 이용하고 있다. 하지만 현재 HDFS의 복제 정책은 모든 데이터에 3개의 복제복을 유지하도록 하고 있다. 하지만 이러한 정책은 데이터의 중요도, 이용빈도수를 반영하지 못한 정책으로 상이한 서비스 품질 및 신뢰성 수준을 반영하지 못한다. 본 논문에서는 Hidden Markov Model을 이용하여 데이터의 이용 빈도수에 따라 복사본의 개수를 조절하는 알고리즘을 제안한다.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2014.01a
• /
• pp.347-348
• /
• 2014

최근 급증하는 교통 혼잡으로 인해 시간적/물질적 손실이 크게 발생하고 있다. 이러한 교통난 해소는 시설투자만으로는 근본적인 해결책이 될 수 없다는 판단 하에 지난 수년간 보다 정확한 교통량을 예측하기 위해 시계열 기반의 다양한 교통량 예측 모델들이 개발 되어 왔다. 그러나 시계열 기반의 모델들은 회귀분석을 통해 과거 교통량을 분석하고 과거의 교통패턴이 미래에도 지속적으로 연장된다는 가정 하에 연구되었기 때문에 실시간으로 급변하는 불규칙한 교통 패턴에 대한 예측의 신뢰성을 떨어트린다. 또한 시계열 기반의 예측 기법은 어떠한 회귀분석 모델을 사용하는지에 따라 성능의 차이가 많이 나타나기 때문에 회귀분석 모델 선택이 중요하다. 이러한 제약을 극복하기 위해 본 논문에서는 은닉 마르코프 모델(Hidden Markov model)을 이용해 동적인 교통 패턴에 따라 현재 상황에 맞는 회귀분석 모델을 선택하는 신뢰도 높은 교통량 예측 시스템을 제안한다.

• Speech Sciences
• /
• v.7 no.4
• /
• pp.11-26
• /
• 2000

In this paper, we propose the HMM with the MIN module. Because initial and re-estimated variance vectors are important elements for performance in HMM recognition systems, we propose a method which compensates for the mismatched statistical feature of training and test data. The MIN module function is a differentiable function similar to the sigmoid function. Unlike a continuous density function, it does not include variance vectors of the data set. The proposed hybrid HMM/MIN module is a unified network in which the observation probability in the HMM is replaced by the MIN module neural network. The parameters in the unified network are re-estimated by the gradient descent method for the Maximum Likelihood (ML) criterion. In estimating parameters, the variance vector is not estimated because there is no variance element in the MIN module function. The experiment was performed to compare the performance of the proposed HMM and the conventional HMM. The experiment measured an isolated number for speaker independent recognition.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.7 no.2
• /
• pp.182-187
• /
• 2006

This paper proposes a method for boundary recognition of named entity using hidden markov model and ontology information of biological named entity. We uses smoothing method using 31 feature information of word and hierarchical information to alleviate sparse data problem in HMM. The GENIA corpus version 2.1 was used to train and to experiment the proposed boundary recognition system. The experimental results show that the proposed system outperform the previous system which did not use ontology information of hierarchical information and smoothing technique. Also the system shows improvement of execution time of boundary recognition.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.15 no.3
• /
• pp.145-152
• /
• 2015

We consider an online selective-sample learning problem for sequence classification, where the goal is to learn a predictive model using a stream of data samples whose class labels can be selectively queried by the algorithm. Given that there is a limit to the total number of queries permitted, the key issue is choosing the most informative and salient samples for their class labels to be queried. Recently, several aggressive selective-sample algorithms have been proposed under a linear model for static (non-sequential) binary classification. We extend the idea to hidden Markov models for multi-class sequence classification by introducing reasonable measures for the novelty and prediction confidence of the incoming sample with respect to the current model, on which the query decision is based. For several sequence classification datasets/tasks in online learning setups, we demonstrate the effectiveness of the proposed approach.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2020.06a
• /
• pp.111-111
• /
• 2020

하천환경과 기후의 변화로 인해 수질오염 과정의 메커니즘이 더욱 복잡해짐에 따라 다양한 요인을 고려한 불확실성 평가 연구가 요구되고 있다. 하천 수질 중에서도 부영양화 문제는 특히 개발로 인한 하천환경 변화 이후 사회 정치적 논점이 되어왔다. 본 연구에서는 지난 7년 동안의 수질 변화의 전반적인 양상을 조사하였으며, 클로로필-a(Chl-a, chlorophyll-a) 농도의 시공간적 의존성의 효과적으로 고려하기 위해 기계학습 기반 분류(classification) 접근법인 다변량 은닉 마코프 모형(MHMM, multivariate hidden Markov model)을 사용하였다. 월 단위 수질 및 수문 자료를 사용하여 Chl-a의 변동성을 군집화하여 수질 상태의 익월 천이확률을 효과적으로 추정하였다. Chl-a와 수질 및 수문기상 조건의 관계를 평가하였으며, 결과적으로 수질 상태의 시공간적 전이가 정확하게 식별되었고 이의 잠재적 원인에 대하여 논의하였다.

• ETRI Journal
• /
• v.46 no.2
• /
• pp.250-262
• /
• 2024

Air-writing recognition is relevant in areas such as natural human-computer interaction, augmented reality, and virtual reality. A trajectory is the most natural way to represent air writing. We analyze the recognition accuracy of words written in air considering five features, namely, writing direction, curvature, trajectory, orthocenter, and ellipsoid, as well as different parameters of a hidden Markov model classifier. Experiments were performed on two representative datasets, whose sample trajectories were collected using a Leap Motion Controller from a fingertip performing air writing. Dataset D₁ contains 840 English words from 21 classes, and dataset D₂ contains 1600 English words from 40 classes. A genetic algorithm was combined with a hidden Markov model classifier to obtain the best subset of features. Combination ftrajectory, orthocenter, writing direction, curvatureg provided the best feature set, achieving recognition accuracies on datasets D₁ and D₂ of 98.81% and 83.58%, respectively.

• Journal of Korea Water Resources Association
• /
• v.46 no.10
• /
• pp.1029-1040
• /
• 2013

This study aims to develop a multivariate daily rainfall simulation model considering spatial coherence across watershed. The existing Hidden Markov Model (HMM) has been mainly applied to single site case so that the spatial coherences are not properly addressed. In this regard, HMM coupled with Chow-Liu Tree (CLT) that is designed to consider inter-dependences across rainfall networks was proposed. The proposed approach is applied to Han-River watershed where long-term and reliable hydrologic data is available, and a rigorous validation is finally conducted to verify the model's capability. It was found that the proposed model showed better performance in terms of reproducing daily rainfall statistics as well as seasonal rainfall statistics. Also, correlation matrix across stations for observation and simulation was compared and examined. It was confirmed that the spatial coherence was well reproduced via CLT-HMM model.

• Electronics and Telecommunications Trends
• /
• v.5 no.2
• /
• pp.63-77
• /
• 1990

HMM(Hidden Markov Model)은 음성을 기술하는데 적합한 model이다. 본 고는 최근 CMU에서 개발한 HMM에 기반을 둔 화자독립, 연속음성 system인 SPIHNX에 대하여 기술한다. SPHINX는 단순한 음소의 HMM model을 적용한 baseline SPHINX로부터 시작하여 새로운 지식의 추가 및 음성단위의 조정 등을 통하여 지속적으로 그 성능이 개선되어 왔다. SPHINX의 최종 version은 어휘 약 1000단어 정도의 재원 관리에 관한 질문 형태의 문장을 인식하는데 96%의 높은 인식율을 보인다. SPHINX는 가장 발전된 음성인식 시스템의 하나이며 이는 화자독립, 대용량어휘의 연속음성 인식 시스템의 실현 가능성을 제시한다.

• Asian Pacific Journal of Cancer Prevention
• /
• v.15 no.9
• /
• pp.4109-4115
• /
• 2014

Background: Accurate assessment of disease progression requires proper understanding of natural disease process which is often hidden and unobservable. For this purpose, disease status should be clearly detected. But in most diseases it is not possible to detect such status. This study, therefore, aims to present a model which both investigates the unobservable disease process and considers the error probability in diagnosis of disease states. Materials and Methods: Data from 330 patients with gastric cancer undergoing surgery at the Iran Cancer Institute from 1995 to 1999 were analyzed. Moreover, to estimate and assess the effect of demographic, diagnostic and clinical factors as well as medical and post-surgical variables on transition rates and the probability of misdiagnosis of relapse, a hidden Markov multi-state model was employed. Results: Classification errors of patients in alive state without a relapse ($e_{21}$) and with a relapse ($e_{12}$) were 0.22 (95% CI: 0.04-0.63) and 0.02 (95% CI: 0.00-0.09), respectively. Only variables of age and number of renewed treatments affected misdiagnosis of relapse. In addition, patient age and distant metastasis were among factors affecting the occurrence of relapse (state1${\rightarrow}$state2) while the number of renewed treatments and the type and extent of surgery had a significant effect on death hazard without relapse (state2${\rightarrow}$state3)and death hazard with relapse (state2${\rightarrow}$state3). Conclusions: A hidden Markov multi-state model provides the possibility of estimating classification error between different states of disease. Moreover, based on this model, factors affecting the probability of this error can be identified and researchers can be helped with understanding the mechanisms of classification error.