• 한국음향학회지
• /
• 제21권7호
• /
• pp.656-664
• /
• 2002

본 논문에서는 한국어 발음열 자동 생성기를 이용하여 한국어의 음운 규칙에 대한 통계적 분석을 수행하였다. 실험에 사용한 발음열 자동 생성기는 한국어 음운 변화 현상에 대해 형태음운론에 기반 한 언어학적 분석과 문교부 표준어 규정의 표준 발음법에서 유도된 필수 및 수의적 음소 변동 규칙과 변이음 규칙의 단계적 적용 모델을 사용해서 구현되었으며, 특히 연속음성 인식을 위한 학습용 발음열과 인식용 발음사전 생성의 최적화를 목표로 하였다. 본 논문에서는 대어휘 연속음성 인식기의 음향 모델을 구축하기 위해 만들어진 삼성 PBS(Phonetically Balanced Sentence) 음성 데이터 베이스의 60,000문장에 적용된 발음열 생성기의 음소 변동규칙들의 분포 및 그 통계를 사용해서 한국어 음운 변화 양상을 분석하였다. 적용된 빈도수를 기준으로 분석한 결과, 필수음소 변동규칙의 경우는 연음법칙, 경음화, 격음화, 장애음의 비음화순으로, 수의적 음소 변동규칙의 경우는 초성 ㅎ 탈락, 중복 자음화, 동일 조음위치 자음탈락 순으로 음운 변화가 발생하였다. 이러한 적용 규칙들의 통계적 자료를 기반으로 한국어 음운 변화 양상을 파악할 수 있었으며, 나아가 본 논문의 연구 결과는 음성 인식 시스템을 개발하는데 유용하게 사용할 수 있을 것이다.

• 한국소음진동공학회:학술대회논문집
• /
• 한국소음진동공학회 2004년도 추계학술대회논문집
• /
• pp.155-159
• /
• 2004

In this paper, it was analyzed the characteristics of electrical and acoustical variations for loudspeaker due to fabrication processes. First, mass of each components of loudspeaker was measured by electric precision scale and performed statistical analysis. Second. Thiele-Small parameters of sample loudspeakers produced by unskilled students were identified by known mass parameter identification method using electrical impedance method and investigated on the variations of each parameter. Electrical impedance tests and acoustic frequency responses were measured on sample loudspeakers and variations were examined to grasp relationship between components variation and fabrication processes. Main factors to effect the changes of electrical impedance were concluded by fabrication processes errors not by components of loudspeaker.

• Communications for Statistical Applications and Methods
• /
• 제27권2호
• /
• pp.201-210
• /
• 2020

Baseline-category logit random effects models have been used to analyze longitudinal nominal data. The models account for subject-specific variations using random effects. However, the random effects covariance matrix in the models needs to explain subject-specific variations as well as serial correlations for nominal outcomes. In order to satisfy them, the covariance matrix must be heterogeneous and high-dimensional. However, it is difficult to estimate the random effects covariance matrix due to its high dimensionality and positive-definiteness. In this paper, we exploit the modified Cholesky decomposition to estimate the high-dimensional heterogeneous random effects covariance matrix. Bayesian methodology is proposed to estimate parameters of interest. The proposed methods are illustrated with real data from the McKinney Homeless Research Project.

• Asian-Australasian Journal of Animal Sciences
• /
• 제31권9호
• /
• pp.1381-1386
• /
• 2018

Animal experiments are essential to the study of animal nutrition. Because of the large variations among individual animals and ethical and economic constraints, experimental designs and statistical analyses are particularly important in animal experiments. To increase the scientific validity of the results and maximize the knowledge gained from animal experiments, each experiment should be appropriately designed, and the observations need to be correctly analyzed and transparently reported. There are many experimental designs and statistical methods. This editorial does not aim to review and present particular experimental designs and statistical methods. Instead, we discuss some essential elements when designing an animal experiment and conducting statistical analyses in animal nutritional studies and provide guidelines for submitting a manuscript to the Asian-Australasian Journal of Animal Sciences for consideration for publication.

• JSTS:Journal of Semiconductor Technology and Science
• /
• 제14권6호
• /
• pp.824-831
• /
• 2014

In this study, structural variations and overlay errors caused by multiple patterning lithography techniques to print narrow parallel metal interconnects are investigated. Resistance and capacitance parasitic of the six lines of parallel interconnects printed by double patterning lithography (DPL) and triple patterning lithography (TPL) are extracted from a field solver. Wide parameter variations both in DPL and TPL processes are analyzed to determine the impact on signal propagation. Simulations of 10% parameter variations in metal lines show delay variations up to 20% and 30% in DPL and TPL, respectively. Monte Carlo statistical analysis shows that the TPL process results in 21% larger standard variation in delay than the DPL process. Crosstalk simulations are conducted to analyze the dependency on the conditions of the neighboring wires. As expected, opposite signal transitions in the neighboring wires significantly degrade the speed of signal propagation, and the impact becomes larger in the C-worst metals patterned by the TPL process compared to those patterned by the DPL process. As a result, both DPL and TPL result in large variations in parasitic and delay. Therefore, an accurate understanding of variations in the interconnect parameters by multiple patterning lithography and adding proper margins in the circuit designs is necessary.

• 한국우주과학회:학술대회논문집(한국우주과학회보)
• /
• 한국우주과학회 2004년도 한국우주과학회보 제13권1호
• /
• pp.36-36
• /
• 2004

Ionospheric data from DGS-256 ionosonde operated by Radio Research Laboratory in Anyang archived during 1991-2002 was extracted and analyzed firstly in Korea. Daily, monthly and annual variations of the 12-year F2 layer critical frequency(foF2) are derived to investigate the statistical ionospheric characteristics during one complete solar cycle. Positive correlation between the mean values of 24-hourly monthly median foF2 and the monthly smoothed sunspot number(SSN) for the same period is found. (omitted)

• 응용통계연구
• /
• 제24권1호
• /
• pp.103-114
• /
• 2011

시장 간 통합을 검정한 연구들은 시장변수들 자체의 연관성으로만 분석을 한정시키는 경향이 있어 경우에 따라 시장 변동체계의 전반적 현실을 파악하는데 한계가 있다. 주식시장의 경우 위기를 겪은 나라와 그렇지 않은 나라와는 일정 기간 주가변동 성향이 다르므로 이들의 동적 연관성에 대한 연구에 선행연구들과 같이 주가만 고려할 경우 주식시장에 영향을 미친 변수들을 제외함에 따른 통계적 편의가 존재하게 된다. 본 연구에서는 우리나라와 주요 투자국의 주식시장 간 통계적 통합의 검정모형에 각국의 주가 외에 국내 외환 및 금융시장을 동시에 포함시켜 보았다. 분석 결과 위기에 따른 변화의 영향이 계속되는 기간에는 이들이 주식시장의 통합에 유의한 영향을 미치는 것으로 추정되어 주식시장만 고려할 경우 모형의 설정오류 가능성이 존재함을 입증한다.

• 응용통계연구
• /
• 제24권5호
• /
• pp.847-859
• /
• 2011

cDNA microarray-based comparative genomic hybridization(CGH) data includes low-intensity spots and thus a statistical strategy is needed to detect subtle differences between different cancer classes. In this study, genes displaying a high frequency of alteration in one of the different classes were selected among the pre-selected genes that show relatively large variations between genes compared to total variations. Utilizing copy-number changes of the selected genes, this study suggests a statistical approach to predict patients' classes with increased performance by pre-classifying patients with similar genetic alteration scores. Two-stage logistic regression model(TLRM) was suggested to pre-classify homogeneous patients and predict patients' classes for cancer prediction; a decision tree(DT) was combined with logistic regression on the set of informative genes. TLRM was constructed in cDNA microarray-based CGH data from the Cancer Metastasis Research Center(CMRC) at Yonsei University; it predicted the patients' clinical diagnoses with perfect matches (except for one patient among the high-risk and low-risk classified patients where the performance of predictions is critical due to the high sensitivity and specificity requirements for clinical treatments. Accuracy validated by leave-one-out cross-validation(LOOCV) was 83.3% while other classification methods of CART and DT performed as comparisons showed worse performances than TLRM.

• ETRI Journal
• /
• 제25권5호
• /
• pp.401-406
• /
• 2003

A biometric system determines the identity of a person by measuring physical features that can distinguish that person from others. Since biometric features have many variations and can be easily corrupted by noises and deformations, it is necessary to apply machine learning techniques to treat the data. When applying the conventional machine learning methods in designing a specific biometric system, however, one first runs into the difficulty of collecting sufficient data for each person to be registered to the system. In addition, there can be an almost infinite number of variations of non-registered data. Therefore, it is difficult to analyze and predict the distributional properties of real data that are essential for the system to deal with in practical applications. These difficulties require a new framework of identification and verification that is appropriate and efficient for the specific situations of biometric systems. As a preliminary solution, this paper proposes a simple but theoretically well-defined method based on a statistical test theory. Our computational experiments on real-world data show that the proposed method has potential for coping with the actual difficulties in biometrics.

• 대한기계학회논문집A
• /
• 제33권12호
• /
• pp.1449-1454
• /
• 2009

The performance variation of a multibody system is affected by a variation of various design variables of the system. And the effects of design variable variations on the performance variation must be considered in design of a multibody system. Accordingly, a variation analysis of a multibody system needs to be conducted in design of a multibody system. For a variation analysis of a performance, population mean and variance which are called statistical parameters of design variables are needed. However, an evaluation of statistical parameters of design variables is impossible in many practical cases. Therefore, an estimation of statistical parameters of the performance based on sample mean and variance which are called statistic of design variables is needed. In this paper, the variation analysis method for a multibody system based on design variable samples was proposed. And, using the proposed method, a variation analysis of the vehicle ride comfort based on sample statistic of design variables was conducted.