• 한국정보전자통신기술학회논문지
• /
• 제9권4호
• /
• pp.337-342
• /
• 2016

본 연구에서는 수정 선형예측방법으로 목표물의 신호를 정확히 추정하는 방법에 대해서 연구하였다. 선형예측방법은 임의의 안테나 배열소자를 다른 소자들과 선형 결합하여 도래방향 신호를 추정하는 방법이다. 수정 선형예측방법은 최적 가중치와 사후확률방법을 사용하였다. 모의실험을 이용하여 본 연구에서 제안한 방법과 Bartlett 및 MUSIC방법의 성능을 비교 분석하였다. 모의실험조건은 안테나 배열 소자 9개, 목표물 신호 4개[-5o, 0o, 5o, 10o]에서 방향을 추정한다. 모의실험에서 Bartlett과 MUSIC방법은 목표물 신호를 3개만 추정하였고, 본 연구에서 제안한 방법은 목표물 신호 4개를 모두 추정하였다. 본 연구에서 제안한 방법이 기존의 Bartlett 과 MUSIC방법보다 분해능이 우수함을 나타내었다.

• Journal of Periodontal and Implant Science
• /
• 제35권1호
• /
• pp.65-76
• /
• 2005

This retrospective study evaluates the clinical performance of the recently introduced $XiVE^{(R)}$ implant(Dentsply-Friadent) with a new macro-design to improve primary stability. A total of 208 $XiVE^{(R)}$ implants (101 in the maxilla and 107 in the mandible) were placed in 71 patients. The average age of the patients was 49 years. Of the 208 implants, 190 (91.3%) were posterior implants and 82 (39.4%) were placed in compromised sites (grafted sites). Clinical and radiographic evaluation were made at second stage surgery for exposure and after functional loading. 192 implants in 64 patients were evaluated at exposure and 146 implants in 50 patients were loaded (average 170 days-loading) and evaluated after functional loading. Of 192 implants available for evaluation before loading, 3 implants failed (early failure) ; 1 before exposure, 1 at exposure and 1 during prosthetic procedure. 2 implants were in the maxilla and 1 was in the mandible. The success rate before loading was 98.4%. After functional loading, no implant failure was occurred in 146 implants evaluated during this period (100% interval success rate). This preliminary data with a new implant showed excellent success rate although the majority of implants evaluated in this study were placed in the posterior region of the jaw and compromised sites.

• IEIE Transactions on Smart Processing and Computing
• /
• 제4권4호
• /
• pp.202-208
• /
• 2015

In this study, we propose a new inference algorithm for a multiclass Gaussian process classification model using a variational EM framework and the Laplace approximation (LA) technique. This is performed in two steps, called expectation and maximization. First, in the expectation step (E-step), using Bayes' theorem and the LA technique, we derive the approximate posterior distribution of the latent function, indicating the possibility that each observation belongs to a certain class in the Gaussian process classification model. In the maximization step, we compute the maximum likelihood estimators for hyper-parameters of a covariance matrix necessary to define the prior distribution of the latent function by using the posterior distribution derived in the E-step. These steps iteratively repeat until a convergence condition is satisfied. Moreover, we conducted the experiments by using synthetic data and Iris data in order to verify the performance of the proposed algorithm. Experimental results reveal that the proposed algorithm shows good performance on these datasets.

• 대한전자공학회논문지SP
• /
• 제44권1호
• /
• pp.40-48
• /
• 2007

이 논문은 다중 스케일 베이지안 관점에서 다층 퍼셉트론과 마코프 랜덤 필드를 사용한 새로운 결 분할 방법을 제안한다. 다층 퍼셉트론의 출력은 사후 확률을 모델링하므로 본 논문에서는 다중 스케일 웨이블릿 계수들을 다층 퍼셉트론의 입력으로 사용한다. 다층 퍼셉트론으로부터 구한 사후 확률과 MAP (maximum a posterior) 분류를 이용하여 각 스케일에서 결 분류를 수행한다. 또한 가장 섬세한 스케일에서 더 개선된 분할 결과를 얻기 위하여 모든 스케일에서 MAP 분류 결과들을 거친 스케일에서 섬세한 스케일까지 차례로 융합한다. 이런 과정은 한 스케일에서의 분류 정보와 그 인접한 보다 거친 스케일에서 얻어지는 문맥과 관련한 연역적 정보를 이용하여 MAP 분류를 행함으로써 이루어진다. 이 융합 과정에서, MRF (Markov random fields) 사전 모델이 평탄화 제한자로서 동작하고, 깁스 샘플러 (Gibbs sampler)는 MAP 분류기로서 동작한다. 제안한 분할 방법은 HMT (Hidden Markov Trees) 모델과 HMTseg 알고리즘을 이용한 결 분할 방법보다 더 좋은 성능을 보인다.

• 대한물리치료과학회지
• /
• 제10권1호
• /
• pp.109-121
• /
• 2003

The purpose of this study was to test the difference of sitting balance control between a normal group and a group of patients with low back pain when their eyes were opened or closed. The 30 subjects of the control group had been chosen from healthy individuals who fit into the pre-designed criteria, and the 30 subjects of the experimental group were composed of the patients with LBP who had their treatment from S hospital from september 1, 2002, to October 30, 2002, and the subjects were measured by static balance test by using a balance performance monitor(BPM). Static balance test was done twice for each subject with his or her eyes opened and closed. Collected data were statistically analyzed by SPSS/PC using unpaired T-Test, Pained T-Test and multiple regression. The results were as follows: 1. In static balance test, normal group did not show statistical significance in sway angle(Anterior, Posterior, left and Right), sway path, sway area and maximal sway velocity, but showed statistical significance in mean balance with eyes opened and eyes dosed(P<.05) 2. In static balance test, LBP group did not show statistical significance in sway angle(Anterior, Posterior, left and Right), mean balance, sway path, sway area and maximal sway velocity with eyes opened and eyes dosed 3. With eyes opened, the comparison between the normal group and the LBP group showed statistical significance in sway angle(Anterior, Posterior, left and Right), mean valance, sway path, sway area and maximal sway velocity(p<.05). With eyes closed, normal group and LBP group did not show statistical significance in sway angle(Anterior and Right), sway area, but showed statistical significance in sway angle(Posterior and Left), mean balance, sway path, sway area and maximal sway velocity(p<.05) In conclusion, there was a significant difference in static sitting balance between normal group and LBP patients group. For future studies, I strongly suggest that researches be done on the treatment with LBP by predicting changes of postures and manipulating them.

• 응용통계연구
• /
• 제11권1호
• /
• pp.97-111
• /
• 1998

본 논문에서는 회귀모형 오차항의 1차 자기상관에 대한 베이즈 검정법을 제안하였다. 이를 위해 자기상관검정에서 설정된 귀무 및 대립가설간에 베이즈 요인을 도출하고, 이를 근사추정하는 방법을 일반화 Savage-Dickey 밀도비와 Gibbs 추출법의 합성을 통해 제시하였다. 또한, 근사추정의 효율 및 제안된 검정법의 검정력을 평가하기 위해서 모의실험과 경험적 자료분석 예를 사용하였다.

• 응용통계연구
• /
• 제21권5호
• /
• pp.835-843
• /
• 2008

In this paper, we investigate a Bayesian inference for software reliability models based on mean value functions which take the form of the mixture of beta distribution functions. The posterior simulation via the Markov chain Monte Carlo approach is used to produce estimates of posterior properties. Its applicability is illustrated with two real data sets. We compute the predictive distribution and the marginal likelihood of various models to compare the performance of them. The model comparison results show that the model based on the beta-mixture performs better than other models.

• 말소리와 음성과학
• /
• 제6권3호
• /
• pp.155-164
• /
• 2014

This paper proposes a new method to determine the recognition units for large vocabulary continuous speech recognition (LVCSR) in Korean by applying unsupervised segmentation and merging. In the proposed method, a text sentence is segmented into morphemes and position information is added to morphemes. Then submorpheme units are obtained by splitting the morpheme units through the maximization of posterior probability terms. The posterior probability terms are computed from the morpheme frequency distribution, the morpheme length distribution, and the morpheme frequency-of-frequency distribution. Finally, the recognition units are obtained by sequentially merging the submorpheme pair with the highest frequency. Computer experiments are conducted using a Korean LVCSR with a 100k word vocabulary and a trigram language model obtained by a 300 million eojeol (word phrase) corpus. The proposed method is shown to reduce the out-of-vocabulary rate to 1.8% and reduce the syllable error rate relatively by 14.0%.

• 한국멀티미디어학회논문지
• /
• 제23권4호
• /
• pp.595-602
• /
• 2020

We propose a generative probabilistic model with Dirichlet prior distribution for topic modeling and text similarity analysis. It assigns a topic and calculates text correlation between documents within a corpus. It also provides posterior probabilities that are assigned to each topic of a document based on the prior distribution in the corpus. We then present a Gibbs sampling algorithm for inference about the posterior distribution and compute text correlation among 50 abstracts from the papers published by IEEE. We also conduct a supervised learning to set a benchmark that justifies the performance of the LDA (Latent Dirichlet Allocation). The experiments show that the accuracy for topic assignment to a certain document is 76% for LDA. The results for supervised learning show the accuracy of 61%, the precision of 93% and the f1-score of 96%. A discussion for experimental results indicates a thorough justification based on probabilities, distributions, evaluation metrics and correlation coefficients with respect to topic assignment.

• 대한의용생체공학회:의공학회지
• /
• 제38권2호
• /
• pp.74-81
• /
• 2017

The purpose of this study was to investigate the EMG characteristics of driver's upper extremity and driving performance for manipulating brake and accelerator pedal by using left and right hand control devices during simulated driving. The people with disabilities in the lower limb have problems in operation of the motor vehicle because of functional loss for manipulating brake and accelerator pedal. Therefore, if hand control device is used for adaptive driving controls in people with lower limb impairments, the disabled people can improve their quality of life by driving a motor vehicle. Six subjects were participated in this study to evaluate driving performance and muscle activities for operating brake and accelerator pedal by using two different hand controls (steering column mounted hand control and floor mounted hand control) in driving simulator. We measured EMG activities of six muscles (posterior deltoid, middle deltoid, triceps, biceps, flexor carpi radialis, and extensor carpi radialis) during pushing and pulling movement with different hand controls for acceleration and braking. STISim Drive 3 software was used for the performance test of different hand control devices in straight lane course for time to reach target speed and brake reaction time. While pulling the hand control lever toward the driver, normalized EMG activities of middle deltoid, triceps and flexor carpi radialis in subjects with disabilities were significantly increased (p < 0.05) compared to the normal subjects. It was also found that muscle responses of posterior deltoid were significantly increased (p < 0.05) when using the right hand control than left hand control. While pushing the hand control lever forward away from the driver, normalized EMG activities of posterior deltoid, middle deltoid and extensor carpi radialis in subjects with disability were significantly increased (p < 0.05) compared to the normal subjects. It was shown that muscle responses of middle deltoid, biceps and extensor carpi radialis were significantly increased when using the right hand control than left hand control. Brake reaction time and time to reach target speed in subjects with disability was increased by 12% and 11.3% on average compared to normal subjects. The subjects with physical disabilities showed a tendency to relatively slow acceleration at the straight lane course.