• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제8권3호
• /
• pp.109-114
• /
• 2019

기존 표절 탐지 시스템은 형태소 분석을 기반으로 공통 단어의 빈도수를 이용해 문서의 유사도를 측정한다. 그러나 주제가 같아 유사 단어가 많이 쓰인 경우, 문장 단위로 일부만 발췌 표절한 경우, 그리고 조사와 어미의 유사성이 있는 경우는 공통 단어의 빈도수만으로는 정확한 유사도를 측정하는데 한계가 있다. 따라서 본 논문에서는 공통 단어 빈도수 기반의 유사도 측정 외에 문장 유사도와 어절 유사도를 추가적으로 측정해 유사도의 정확성을 높일 수 있는 표절 탐지 시스템을 설계하고 구현하였다. 실험 결과, 문장 유사도를 측정함으로써 문장 단위로 표절이 이루어진 경우를 발견할 수 있었고, 어절 유사도를 추가로 측정함으로써 부분표절이 일어난 경우라도 조사나 어미까지 그대로 사용한 표절의 경우 등을 발견할 수 있었다.

• 대한음성학회지:말소리
• /
• 제37호
• /
• pp.49-71
• /
• 1999

The Purpose of this paper is to gather pairs of heteronyms in modern Korean and transcribe them in the IPA and the Roman alphabet in order to propose that all of them should be differentiated in Hanngul orthography. More than a quarter of the whole Korean vocabulary consists of words with a long vowel and the number of minimal pairs distinguished only by the chroneme reaches nearly ten thousand (i.e. twenty thousand words). The letter h syllable-finally is used here to represent the long vowel in Romanization except the vowel '으‘[?:] which is transcribed by doubling the letter u (i.e. uu). Another factor bringing forth lots of heteronyms in Korean is the lack of full indication as to the non-automatic reinforcement in the initial consonant of a word (or a morpheme) when preceded by another within a phrase (or a word). These reinforced word-initial consonants are written with the letter c and an apostrophe (like c'g- , c'd- , c'b-, c's-, c'j-) in Romanization here. The reinforced morpheme-initial consonant within a word is written with the letters k t, p, ss and cz for ㄲ, ㄸ, ㅃ, ㅆ and ㅉ sounds respectively. The contrasted pronunciations of pairs of heteronyms beginning with ㅅ /s/sup h// and ㅆ /s/ sounds are transcribed here for exemplification.

• 패션비즈니스
• /
• 제10권1호
• /
• pp.106-119
• /
• 2006

The purpose of this study was to analysis how perceived beauty service quality, perceived value influenced on purchase intention of customers. Moreover, this study was suggested strategic solutions for beauty managers and employees to build efficient marketing strategies in order to maximize customer satisfaction. For research method, SPSS 12.0 program was used and for the accuracy of precision of a measuring instrument reliability test was applied. And then simple and multiple regression analysis were performed to examine the effects of independent variables to dependent variables to dependant variables. The results of this study were five folds: the higher perceived service quality of beauty service was, the higher perceived value, satisfaction, word-of-mouth, and intention of revisiting were; the higher perceived value was, the better satisfaction, word-of-mouth, and intention of revisiting were; technical service quality was single most important factor affecting influence on perceived value, satisfaction, word of mouth, intention of revisiting, and functional service quality and physical service quality were listed second and third.

• 말소리와 음성과학
• /
• 제7권1호
• /
• pp.107-116
• /
• 2015

The present experiment investigated the effect of word length on the length of f0 intervals for North Kyungsang children. In order to find out the lengths of the f0 intervals, the f0 values at the midpoints of vowels in words were measured. F0 estimates were computed as intervals consistent with the logarithmic scale corresponding to the number of syllables in the words. The results indicated that the mean f0 intervals in words of different lengths showed a significant difference for the HH in HH vs. HHL and the LH in LH vs. LLH for North Kyungsang children. Adult speakers from the North Kyungsang region significantly differed only within the HH in HH vs. HHL. Adult speakers made a noticeable contribution in this characteristic from the children. The result of the adult study was presented to confirm whether the children used a North Kyungsang dialect. With respect to individual speaker differences, the North Kyungsang children showed more or less consistent patterns in quantile-quantile plots for the HH vs. HHL, but for the HL vs. LHL and LH vs. LLH, there were more variations than for the HH vs. HHL. The individual speakers' variation was the largest for the HL vs. LHL and the smallest for HH vs. HHL. Considering these results, the effect of word length on f0 intervals tended to show pitch accent-type-specific characteristics in the process of prosodic acquisition.

• 한국통신학회논문지
• /
• 제17권1호
• /
• pp.20-28
• /
• 1992

조음결합은 연결단어 음성인식에서 오인식을 발생시키는 주된 요인이다. 본 연구에서는 표준패턴의 접속 부분에서 선행단어의 영향을 받아 후속단어의 선두 부분하는 것을 고려해 조음결합 모델을 적용하여 후속단어에 변형을 주었다. 조음결합 모델은 임계제동형 선형 2차계를 이용하였으며 1단 DP매칭 인식 알고리즘에 이 모델을 조합하여 효과를 조사했다. 그 결과 화자 5명에 의한 4연 숫자음 35어의 인식 실험에서 조음결합 처리에 의한 인식율이 평균 4.7%개선 되었다.

• 패션비즈니스
• /
• 제13권4호
• /
• pp.51-59
• /
• 2009

The purpose of this research is to investigate the influences of service experience and brand reputation on intention by word of mouth(WOM) when consumers get experiences at beauty salon. We made 2(valence of service experience: positive/negative) ${\times}$ 2(brand reputation level: high/low) subjects experimental design in order to find reason that there is a significant difference of intention by word of the mouth between positive experience condition and negative experience condition. Also we tried to find the interaction of two factors. To test our hypothesis, an experiment was performed on 68 university students in seoul area. Specifically, we used 2-way ANOVA and planned contrast method through SPSS 12.0 procedure. As a result, a two-way ANOVA demonstrated a significant main effect of service experience, which means that people more intend to make WOM when they experience a negative service than a positive service. Also the ANOVA showed a significant interaction effect between service experience and brand reputation, which represents that people more intend to make WOM for high-reputed(low-reputed) brand than low-reputed(high-reputed) brand when they encounter negative(positive) services.

• 대한한의학원전학회지
• /
• 제32권3호
• /
• pp.47-57
• /
• 2019

Objectives : We aim to assist in choosing an appropriate method for word extraction when analyzing East Asian Traditional Medical texts based on unsupervised learning. Methods : In order to assign ranks to substrings, we conducted a test using one method(BE:Branching Entropy) for exterior boundary value, three methods(CS:cohesion score, TS:t-score, SL:simple-ll) for interior boundary value, and six methods(BExSL, BExTS, BExCS, CSxTS, CSxSL, TSxSL) from combining them. Results : When Miss Rate(MR) was used as the criterion, the error was minimal when the TS and SL were used together, while the error was maximum when CS was used alone. When number of segmented texts was applied as weight value, the results were the best in the case of SL, and the worst in the case of BE alone. Conclusions : Unsupervised-Learning-Based Word Extraction is a method that can be used to analyze texts without a prepared set of vocabulary data. When using this method, SL or the combination of SL and TS could be considered primarily.

• 한국산학기술학회논문지
• /
• 제10권7호
• /
• pp.1680-1684
• /
• 2009

한글 단어 재인에 있어서 음절 내 단위가 음절체인지에 관한 실험증거를 제공하기 위하여 점화과제를 사용하였다. 실험조건으로 음절체가 중복되는 조건 (예: 섬질 -> 성직), 각운이 중복되는 조건 (예: 헝칙 -> 성직), 초성과 종성이 중복되는 조건 (예: 생적 -> 성직), 그리고 아무런 철자도 중복되지 않는 조건 (예: 찬멸 -> 성직)를 서로 비교하였다. 또한 목표자극의 글자 유형을 <각> 형과 <곡> 형으로 구분하여 차별적인 점화효과가 일어나는지를 알아보았다. 실험 결과, <각> 형의 목표자극에서만 유의미한 억제 점화 효과가 나타났다. 이는 특정 유형의 단어에 한하여 한글의 단어 재인의 단위가 음절체 일 수 있음을 시사하는 결과이며 보다 다양한 글자 유형에 대하여 점화 효과를 알아보는 연구가 필요하다.

• 음성과학
• /
• 제8권4호
• /
• pp.275-288
• /
• 2001

The sentences with the same surface word order may be realized with the pragmatically different meanings, depending on the contexts under which they could appear. Semantically, their meaning differences have been explained in terms of the different information structures (Steedman 2000), whereas prosodically, they can be explained in terms of the different compositions of intonational components which make their own semantic contributions (Pierrehumbert and Hirschberg 1990). In other words, the different intonation realizations of the sentences with the same word order reflect the different information structures. In this paper, we investigate the relationship between the information structure and the intonational meaning by way of analysing the production of the sentences with ambiguous scopes of the English focus particle 'only'. In contrast to the previous quantitative approaches to the scopes of the focus particle 'only', two independent levels of information structure (Steedman 2000)-theme/rheme, and focus/ background-make it possible to consistently explain the intonational phenomena.

• 디지털융복합연구
• /
• 제18권1호
• /
• pp.301-307
• /
• 2020

최근 추천 시스템은 5G 시대의 시작과 동시에 여러 분야에서 도입하고 있으며, 주로 도서나 영화, 음악 분야의 서비스에서 크게 두각을 나타내고 있다. 그러나 이러한 추천 시스템에서 사용자마다 선호하는 정도가 주관적이고, 불확실하여 정확한 추천 서비스를 제공하기가 어렵다. 추천 시스템의 성능을 향상시키기 위해서는 많은 양의 학습 데이터가 필요하며, 추론 기술이 보다 정확해야 한다. 이러한 문제점을 해결하기 위하여 본 연구에서는 Fuzzy-AHP와 Word2Vec 학습 기법을 이용한 영화 추천 시스템을 제안하였다. 본 시스템에서는 사용자의 선호도를 객관적으로 예측하기 위해 Fuzzy-AHP를 사용하였으며, 스크레이핑한 데이터를 분류하기 위해 Word2Vec 학습 기법을 사용하였다. 본 시스템의 성능을 평가하기 위해 그리드 서치를 이용하여 Word2Vec 학습 결과의 정확도를 측정하였고, 그 후 본 시스템이 예측한 평점과 관객들이 평가한 영화의 평점 간 차이를 비교하였다. 그 결과 최적의 교차 검증 정확도가 91.4%로 우수한 성능을 나타내었으며, 예측한 평점과 관객들이 평가한 영화의 평점 간 차이를 Fuzzy-AHP 시스템과 비교한 결과 10% 정도 우수함을 확인할 수 있었다.