• 음성과학
• /
• 제13권2호
• /
• pp.75-84
• /
• 2006

Various methods have been proposed to overcome the problem of speech recognition in the noisy conditions. Among them, the model compensation methods like the parallel model combination (PMC) and Jacobian adaptation (JA) have been found to perform efficiently. The JA is quite effective when we have hidden Markov models (HMMs) already trained in a similar condition as the target environment. In a previous work, we have proposed an improved method for the JA to make it more robust against the changing environments in recognition. In this paper, we further improved its performance by compensating the delta-mean vectors and covariance matrices of the HMM and investigated its feasibility in the multi-model structure for the noisy speech recognition. From the experimental results, we could find that the proposed improved the robustness of the JA and the multi-model approach could be a viable solution in the noisy speech recognition.

• Journal of the Korean Statistical Society
• /
• 제26권4호
• /
• pp.477-491
• /
• 1997

A new predictive classification rule for assigning future cases into one of two multivariate normal population (with unknown normal mixture model) is considered. The development involves calculation of posterior probability of each possible normal-mixture model via a default Bayesian test criterion, called intrinsic Bayes factor, and suggests predictive distribution for future cases to be classified that accounts for model uncertainty by weighting the effect of each model by its posterior probabiliy. In this paper, our interest is focused on constructing the classification rule that takes care of uncertainty about the types of covariance matrices (homogeneity/heterogeneity) involved in the model. For the constructed rule, a Monte Carlo simulation study demonstrates routine application and notes benefits over traditional predictive calssification rule by Geisser (1982).

• Communications for Statistical Applications and Methods
• /
• 제20권1호
• /
• pp.63-75
• /
• 2013

In this paper, we derive maximum likelihood estimators (MLEs) and approximate maximum likelihood estimators (AMLEs) of unknown parameters in a generalized half logistic distribution under Type-II hybrid censoring. We also obtain approximate confidence intervals using asymptotic variance and covariance matrices based on the MLEs and the AMLEs. As an illustration, we examine the validity of the proposed estimation using real data. Finally, we compare the proposed estimators in the sense of the mean squared error (MSE), bias, and length of the approximate confidence interval through a Monte Carlo simulation for various censoring schemes.

• 음성과학
• /
• 제10권4호
• /
• pp.149-157
• /
• 2003

In this paper, we proposed the principal component analysis (PCA) fuzzy mixture model for speaker identification. A PCA fuzzy mixture model is derived from the combination of the PCA and the fuzzy version of mixture model with diagonal covariance matrices. In this method, the feature vectors are first transformed by each speaker's PCA transformation matrix to reduce the correlation among the elements. Then, the fuzzy mixture model for speaker is obtained from these transformed feature vectors with reduced dimensions. The orthogonal Gaussian Mixture Model (GMM) can be derived as a special case of PCA fuzzy mixture model. In our experiments, with having the number of mixtures equal, the proposed method requires less training time and less storage as well as shows better speaker identification rate compared to the conventional GMM. Also, the proposed one shows equal or better identification performance than the orthogonal GMM does.

• ETRI Journal
• /
• 제29권5호
• /
• pp.688-690
• /
• 2007

This letter presents a filter for a car navigation system integrating a low-grade inertial measurements unit (IMU) and a global positioning system receiver. The filter is designed according to the state variables to be estimated and the usable measurements. The usable measurements change from case to case, and the estimative state variables also change due to the measurements; therefore, multiple models must be used for real environmental maneuvers. In this letter, four models for land navigation are chosen and switched by rearranging the system matrix and resetting the error covariance matrices.

• Journal of the Korean Statistical Society
• /
• 제33권1호
• /
• pp.1-24
• /
• 2004

This paper is concerned with statistical methods for multivariate data when the number p of variables is large compared to the sample size n. Such data appear typically in analysis of DNA microarrays, curve data, financial data, etc. However, there is little statistical theory for high dimensional data. On the other hand, there are some asymptotic results under the assumption that both and p tend to $\infty$, in some ratio p/n ${\rightarrow}$c. The results suggest that the new asymptotic results are more useful and insightful than the classical large sample asymptotics. The main purpose of this paper is to review some asymptotic results for high dimensional statistics as well as classical statistics under a high dimensional asymptotic framework.

• 한국항해항만학회지
• /
• 제37권4호
• /
• pp.329-335
• /
• 2013

This paper is to develop the position error equations including the attitude errors, the errors of nadir and ship's heading, and the errors of ship's position in the free-gyro positioning and directional system. In doing so, the determination of ship's position by two free gyro vectors was discussed and the algorithmic design of the free-gyro positioning and directional system was introduced briefly. Next, the errors of transformation matrices of the gyro and body frames, i.e. attitude errors, were examined and the attitude equations were also derived. The perturbations of the errors of the nadir angle including ship's heading were investigated in each stage from the sensor of rate of motion of the spin axis to the nadir angle obtained. Finally, the perturbation error equations of ship's position used the nadir angles were derived in the form of a linear error model and the concept of FDOP was also suggested by using covariance of position error.

• International Journal of Aeronautical and Space Sciences
• /
• 제18권3호
• /
• pp.535-544
• /
• 2017

In-flight aircraft engine performance estimation is one of the key techniques for advanced intelligent engine control and in-flight fault detection, isolation and accommodation. This paper detailed the current performance degradation estimation methods, and an improved hybrid Kalman filter via velocity-based LPV (VLPV) framework for these needs is proposed in this paper. Composed of a nonlinear on-board model (NOBM) and VLPV, the filter shows a hybrid architecture. The outputs of NOBM are used for the baseline of the VLPV Kalman filter, while the system performance degradation factors on-line estimated by the measured real system output deviations are fed back to the NOBM for its updating. In addition, the setting of the process and measurement noise covariance matrices' values are also discussed. By applying it to a commercial turbofan engine, simulation results show the efficiency.

• 응용통계연구
• /
• 제6권2호
• /
• pp.409-419
• /
• 1993

본 논문은 요즈음 국내외를 막론하고 통계학계에서 활발히 연구하고 있는 그래프에 의한 통계적 방법의 일부로서 그래프에 의한 판별분석법을 다루고 있다. 기존에 알려진 Sammon 그래프와 MV 그래프를 바탕으로 새로운 대안의 가능성을 소개하고 있으며 그룹의 수가 2인 경우 실제자료와 모의실험을 이용하여 3가지 그래프의 특성을 비교, 분석하고 있다. 새로운 대안이 해결해야할 차원축소 문제와 로버스트 방법에 대한 앞으로의 과제를 간략히 언급하고 있다.

• 한국측량학회지
• /
• 제30권5호
• /
• pp.445-458
• /
• 2012

GPS 기준점 측량은 관측과 데이터 처리를 통해 측지계에 대한 물리적인 표지의 위치를 목표 정확도 범위 이내로 결정하기 위해 실시하며, 이러한 이유로 측량 정확도를 실제와 유사하게 표현하는 것이 매우 중요한 문제이다. 망조정을 통해 산정되는 기준점 성과의 정확도는 사용되는 수학적 모형에 민감한 영향을 받는 추정좌표의 분산-공분산 행렬에 의해 정량적으로 표현된다. 본 연구는 GPS 망조정에 사용되는 함수모형과 통계모형이 기준점 위치 추정과 정확도 표현에 미치는 영향을 연구하여 향후 정확도 표현의 표준화를 위한 기초자료 확보를 위해 실시하였다. 이를 위하여 단일기선해석 다중세션 망조정 이론과 절차와 방법을 실제 관측데이터 처리를 통한 수치적 분석을 병행한 연구를 수행 하였다. 그 결과 절대정확도와 상대정확도를 보다 현실적으로 반영하여 표현하기 위해서는 잔존하는 관측오차의 모형화가 가능한 경험적 통계모형을 사용하는 다점가중제약조정이 GPS 기준점 성과산정을 위한 수학적 모형으로 보다 타당한 것으로 분석되었다.