• Speech Sciences
• /
• v.9 no.4
• /
• pp.105-110
• /
• 2002

This study compared nasalance in low pressure and high pressure consonants and high and low vowels. The subjects were 8 children with cleft palate ranging from 3 years 6 months to 8 years in age. The measurement were taken in terms of the nasalance scores associated with 20 different word stimuli (high consonants+high vowels, high consonants+low vowels, low consonants+high vowels, and low consonants+low vowels) for subjects. The nasalance scores may be affected by high consonants and vowels types, low vowels and consonants types.

• Language and Information
• /
• v.1
• /
• pp.27-65
• /
• 1997

Nouns in the Korean language are subcategorized for various frames(called SUBCAT lists) in much the same way as verbs are. Assuming a monostratal grammar and building on analyses of various 'little elements' as clitics, such as the ones given by No(1991), Chae(1995,1996), and Oh(1991), I delineate the ranges of SUBCAT lists for the Korean verbs and nouns and show that the two word-classes have heavily overlapping frames. Twenty five SUBCAT lists are identified for verbs, and twenty four for nouns, of which twenty three find associated lexical items in both. By the way of justification, I offer analyses of noun--verb collocations in terms of the new five-valued syntactic feature COLLOC along with SUBCAT, which subsume 'light verb' constructions. It is hoped that this work will have given clear syntactic underpinnings to those who are concerned with practical lexicography.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.14 no.1
• /
• pp.1-7
• /
• 2014

This paper presents a novel method for simultaneously and automatically choosing the nonlinear structures of regressors or discriminant functions, as well as the number of terms to include in a rule-based regression model or pattern classifier. Variables are first partitioned into subsets each of which has a linguistic term (called a causal condition) associated with it; fuzzy sets are used to model the terms. Candidate interconnections (causal combinations) of either a term or its complement are formed, where the connecting word is AND which is modeled using the minimum operation. The data establishes which of the candidate causal combinations survive. A novel theoretical result leads to an exponential speedup in establishing this.

• Proceedings of the KIEE Conference
• /
• 2005.07e
• /
• pp.19-21
• /
• 2005

The purpose of power system design is, in a word, to provide a good quality of electric power. The design of reliable industrial and commercial power distribution system is important because of the high costs associated with power outage. Three major factors for realization of the purpose are: (1) To hold system frequency at or very close to a specified nominal value(e.g. 60Hz) by control of frequency-effective power. (2) To maintain the correct value of interchange power between power and local generators. (3) To hold system voltage at or very close to a specified nominal value by control of voltage-reactive power. Within the past decade, numbers of industrial and commercial facilities installed with local generation, large motors or both, are increasing. This means that system stability is of concern to a growing number of industrial plant electrical engineers and consultants.

• Phonetics and Speech Sciences
• /
• v.2 no.4
• /
• pp.29-37
• /
• 2010

Previous studies have investigated a between-category relationship of multiple acoustic cues for a laryngeal contrast by examining the distributions of VOT, f0 and H1-H2. The current study examined within-category correlations between cues comprising stops by Korean- and English-speaking adults and children to understand how children master the internal structure of stop phonation types in two languages. Word-initial stops were collected from about 70 children and 15 adults speaking English and Korean, and were analyzed in terms of VOT, f0 and H1-H2 to compute correlation coefficients. Findings in adults' productions included a gender-differentiated cue-correlation pattern associated with H1-H2 in Korean tense stops and a trading relationship between f0 and VOT in Korean lax and aspirated stops and English voiced and voiceless stops. Children did not necessarily have adult-like cue-correlation patterns even in early-acquired categories, suggesting that the mastery of intra-category structure of phonation type might occur later than inter-category structure.

• Phonetics and Speech Sciences
• /
• v.9 no.2
• /
• pp.53-59
• /
• 2017

The present study examined if individual listeners' perceptual variations were associated with their cognitive characteristics indexed by the Autistic Spectrum Quotient (AQ). This study first investigated the perception of the lexical pitch accent contrast in the Kyungsang Korean currently undergoing a sound change, and then tested if listeners' perceptual variations were correlated with their AQ scores. Eighteen Kyungsang listeners in their 20s participated in the perception experiment where they identified two contrastive accent words for auditory stimuli systematically varying F0 scaling and timing properties; the participants then completed the AQ questionnaire. In the results, the acoustic parameters reporting reduced phonetic differences across accent contrasts for younger Kyungsang generation played a reliable role in perceiving the HH word from HL, suggesting the discrepancy between the perception and the production in the context of sound change. This study also observed that individuals' perceptual variations were negatively correlated with their AQ sub scores. The present findings suggested that the sound change might appear differently between production and perception with a different time course, and deviant percepts could be explained by individuals' cognitive measure.

• Proceedings of the Korean Information Science Society Conference
• /
• 2010.06c
• /
• pp.395-398
• /
• 2010

많은 정보들을 인터넷을 통하여 접할 수 있게 됨에 따라 사용자가 만족하는 결과를 보여주는 것이 검색 엔진의 궁극적인 목표가 되었다. 하지만 방대한 양을 가진 다양한 정보에서 원하는 검색 결과를 검색하는 것은 과거와 현재까지 많은 연구를 통해 많은 시간과 노력이 필요하다는 것이 증명 되었다. 기존의 HITS 알고리즘을 개선하여 링크 제어를 이용한 페이지와 페이지간에 관련성을 높였다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.10
• /
• pp.3498-3512
• /
• 2021

This article describes how events that make up text stories can be represented and extracted. We also address the results from our simple experiment on extracting and clustering events in terms of emotions, under the assumption that different emotional events can be associated with the classified clusters. Each emotion cluster is based on Plutchik's eight basic emotion model, and the attributes of the NLTK-VADER are used for the classification criterion. While comparisons of the results with human raters show less accuracy for certain emotion types, emotion types such as joy and sadness show relatively high accuracy. The evaluation results with NRC Word Emotion Association Lexicon (aka EmoLex) show high accuracy values (more than 90% accuracy in anger, disgust, fear, and surprise), though precision and recall values are relatively low.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2013.11a
• /
• pp.1189-1192
• /
• 2013

신문기사 분석을 통한 연관어 비주얼라이저는 신문 기사의 단어를 추출하여 단어 간 연관도를 분석하여 다양한 그래프로 표현하는 시스템이다. 인터넷 신문사의 뉴스 기사들을 수집하고 형태소 분석을 통해 기사별로 단어의 출현 횟수를 데이터베이스에 저장하고 단어와 단어 간의 연관성을 분석한다. 단어 간 연관성을 측정하기 위한 기준으로 두 단어 간 동일기사에 존재여부, 동일날짜에 존재여부를 이용한다. 이 값을 바탕으로 웹 페이지 상에서 다양한 그래프로 상위 연관성을 가진 단어들을 표현한다. 표현 되는 그래프는 다양한 형태의 그래프로 단어와 단어사이에 연관성을 보다 쉽게 파악 할 수 있다.

• Journal of Information Processing Systems
• /
• v.16 no.6
• /
• pp.1437-1446
• /
• 2020

For the last few years, smart devices have begun to occupy an essential place in the life of children, by allowing them to access a variety of language activities and books. Various studies are being conducted on using smart devices for education. Our study extracts images and texts from kids' book with smart devices and matches the extracted images and texts to create new images that are not represented in these books. The proposed system will enable the use of smart devices as educational media for children. A deep convolutional generative adversarial network (DCGAN) is used for generating a new image. Three steps are involved in training DCGAN. Firstly, images with 11 titles and 1,164 images on ImageNet are learned. Secondly, Tesseract, an optical character recognition engine, is used to extract images and text from kids' book and classify the text using a morpheme analyzer. Thirdly, the classified word class is matched with the latent vector of the image. The learned DCGAN creates an image associated with the text.