• 인지과학
• /
• 제24권4호
• /
• pp.339-363
• /
• 2013

습득연령효과는 연령 초기에 습득된 단어가 후기에 습득된 단어에 비해 효율적으로 처리되는 현상이다. 습득연령은 단어빈도와 함께 단어처리과정의 주요한 변인으로 간주되고 있다. 본 연구는 한국어/한글 단어재인에서 습득연령효과를 검토하였다. 실험 1에서는 단어명명과제와 어휘판단과제에서 습득연령효과를 검토하였다. 그 결과, 과제유형과 습득연령 간의 상호작용을 관찰하였다. 습득연령효과는 단어명명과제에서는 나타나지 않았고 어휘판단과제에서만 유의하게 나타났다. 실험 2에서는 어휘판단과제를 사용하여 습득연령과 단어빈도의 관계를 검토하였다. 그 결과, 습득연령과 단어빈도는 각각 유의한 변인으로 드러났다. 후기습득 단어에 비해 초기습득 단어의 어휘 판단 수행이 우수했으며, 저빈도 단어에 비해 고빈도 단어의 어휘 판단 수행이 우수했다. 두 변인의 독립적 효과는 확인했으나, 둘 간의 상호작용은 없었다. 실험 3에서는 음 변화가 일어나는 단어 조건과 음 변화가 일어나지 않는 단어 조건에서 습득연령효과를 검토하였다. 그 결과, 습득연령효과는 두 조건에서 유사하게 나타났으며 두 조건 간 차이는 없었다. 본 연구 결과에 대해 여러 가설들을 비교 논의하였다.

• 한국의상디자인학회지
• /
• 제24권3호
• /
• pp.1-15
• /
• 2022

The purpose of this study is to compare and analyze the perception of men's fashion before and after the COVID-19 pandemic. TEXTOM allowed the collection of Big Data based on the term 'men's fashion'. As for the data collection periods, Jan. 1, 2018 to Dec. 31, 2019 was set as the pre-COVID-19 era, while Jan. 1, 2020 to Dec. 31, 2021 was set as the post-COVID-19 era. The top 50 words in terms of appearance frequency were extracted from the data. The extracted words were processed using network centrality analysis and CONCOR analysis using Ucinet 6. Research findings were as follows. 1) In the pre-COVID-19 era, the appearance frequency of 'men' was the highest, followed by 'fashion', 'men's fashion', 'brand', 'daily look', 'suit', and 'department store'. These words came up with a high TF-IDF values. Network centrality analysis discovered that 'men', 'fashion', 'men's fashion', 'brand', and 'suit' had a high level of connectivity with other words. CONCOR analysis showed four significant groups: 'fashion item and styles', 'fashion show', 'purchase', and 'collection'. 2) In the post-COVID-19 era, the appearance frequency of 'men' was the highest, followed by 'fashion', 'brand', 'men's fashion', 'discount', 'women', and 'luxury'. These words also displayed high TF-IDF values. Network centrality analysis found that 'fashion', 'men', 'brand', 'men's fashion', and 'discount' had a high level of connectivity with other words. CONCOR analysis showed four significant groups: 'fashion item and style', 'fashion show', 'purchase', and 'situation'. 3) Before the outbreak of the pandemic, men were interested in suits to wear to the office, daily look, and fashion shows in Milan and Paris. They often purchased menswear in multi-brand and open stores. However, they were more interested in sneakers, casual styles, and online fashion shows as social distancing and working from home became common. Most purchased menswear through online platforms.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권4호
• /
• pp.1400-1418
• /
• 2020

In the development of commercial promotion, chatbot is known as one of significant skill by application of natural language processing (NLP). Conventional design methods are using bag-of-words model (BOW) alone based on Google database and other online corpus. For one thing, in the bag-of-words model, the vectors are Irrelevant to one another. Even though this method is friendly to discrete features, it is not conducive to the machine to understand continuous statements due to the loss of the connection between words in the encoded word vector. For other thing, existing methods are used to test in state-of-the-art online corpus but it is hard to apply in real applications such as telemarketing data. In this paper, we propose an improved chatbot design way using hybrid bag-of-words model and skip-gram model based on the real telemarketing data. Specifically, we first collect the real data in the telemarketing field and perform data cleaning and data classification on the constructed corpus. Second, the word representation is adopted hybrid bag-of-words model and skip-gram model. The skip-gram model maps synonyms in the vicinity of vector space. The correlation between words is expressed, so the amount of information contained in the word vector is increased, making up for the shortcomings caused by using bag-of-words model alone. Third, we use the term frequency-inverse document frequency (TF-IDF) weighting method to improve the weight of key words, then output the final word expression. At last, the answer is produced using hybrid retrieval model and generate model. The retrieval model can accurately answer questions in the field. The generate model can supplement the question of answering the open domain, in which the answer to the final reply is completed by long-short term memory (LSTM) training and prediction. Experimental results show which the hybrid word vector expression model can improve the accuracy of the response and the whole system can communicate with humans.

• 지식경영연구
• /
• 제19권4호
• /
• pp.171-185
• /
• 2018

Studies on online review have carried out analysis of the rating and topic as a whole. However, it is necessary to analyze opinions on various dimensions of service quality. This study classifies reviews of healthcare services into service quality dimensions, and proposes a method to identify words that are mainly referred to in each dimension. Service quality was based on the dimensions provided by SERVQUAL, and patient reviews have collected from NHSChoice. The 2,000 sentences sampled were classified into service quality dimension of SERVQUAL and a method of extracting important keywords from sentences by service quality dimension was suggested. The RAKE algorithm is used to extract key words from a single document and an index is considered to consider frequently used words in various documents. Since we need to identify key words in various reviews, we have considered frequency and discrimination (IDF) at the same time, rather than identifying key words based only on the RAKE score. In SERVQUAL dimension, we identified the words that patients mentioned mainly, and also identified the words that patients mainly refer to by review rating.

• 가정과삶의질연구
• /
• 제26권4호
• /
• pp.41-54
• /
• 2008

The purpose of this study was to investigate predictors concerning preschoolers' ability to read words, in terms of their sub-skills of alphabet knowledge, phonological awareness, and phonological processing. Fourteen literacy sub-tests and three types of reading tasks were administered to 289 kindergartners aged 4 to 6 in Busan. The main results are as follows. Sub-skills that predicted reading ability varied with children's age. Irrespective of children's age groups, knowledge of consonant names and digit naming speed commonly explained the reading of real words. In contrast, skills of syllable deletion and phoneme substitution and knowledge of alphabet composition principles were related to only 4-year-olds' reading skills. Exclusively included was digit memory in predicting 5-year-olds' reading abilities, and knowledge of vowel sounds in 6-year-olds' reading skills. The type of reading task also influenced reading ability. A few common variables such as knowledge of consonant names and vowel sounds, digit naming speed, and phoneme substitution skill explained all types of word reading. Syllable counting skills, however, had predictive value only for the reading of real words. Phoneme insertion skills and digit memory had predictive value for the reading of pseudo words and low frequency letters. Likewise, knowledge of consonant sounds and vowel stroke-adding principles were significant only for the reading of low frequency letters.

• 한국컴퓨터정보학회논문지
• /
• 제16권12호
• /
• pp.83-92
• /
• 2011

일반 출판물과는 달리 문서 편집기를 사용하여 작성중에 있는 문서에는 사용자의 실수에 의한 오타 오류가 자주 발생한다. 이와 같은 온라인 문서에서 맞춤법 오류의 다수를 차지하는 사용자의 오타 오류는 대부분 자판을 입력할 때 주위 문자를 잘못 입력하는 경우이다. 통상적인 철자 검사기는 이러한 오류들을 형태소 분석기를 이용하여 검출하고 교정하게 된다. 즉, 입력된 어절에 대해 형태소 분석을 시도하고 분석되지 않은 어절을 철자 오류로 간주하게 된다. 그러나 오타 입력된 어절임에도 불구하고 형태소 분석에 성공한 경우에는 이와 같은 방법으로는 검출이 불가능하다. 본 논문에서는 기존 방법들이 검출하지 못했던 철자 오류들을 검출해 낼 수 있는 방법을 제시한다. 이 방법은 문서 작성자의 오타 입력은 반복하여 입력되지 않는 경향이 있으므로 저빈도로 발생한다는 특성에 기반하여 제안되었다. 저빈도의 어절의 자소 대치를 통해 문서의 특정 구간 내의 다른 단어와 비교하여 오타일 확률이 적은 단어인 자주 나오는 단어와 매칭이 된다면 일단 오류 후보로 가정하는 것이다. 여기에는 몇 가지 경험적인 제약이 추가되어야 한다. 이러한 단어간 비교에 의한 추정은 기존에 발견하지 못했던 구문오류뿐만 아니라 일부 의미오류까지 검출할 수 있으며, 교정 후보 선정시 가중치 적용에도 사용될 수 있다.

• 음성과학
• /
• 제12권3호
• /
• pp.59-77
• /
• 2005

The aim of the present study is to investigate how the phonemic characteristics influence on the disfluencies of children and adults who stutter. The participants were 10 children(9 boys and 1 girl) and 10 male adults. After having the participants to read out the Paradise-Fluency Assessment(Sim, Shin & Lee, 2004) passages, each of the productions were divided into syllables and words, and then the frequencies and the ratios of their disfluenceis were analyzed according to the specified phonemic features. In terms of the frequency of the disfluency, the participants stuttered more in the words which start with consonant than vowel. But they showed more disfluencies in the words initiated with vowel than consonant when the ratio of each phoneme's presences were considered. There found different tendencies among the phonemic features related with their disfluencies occuring with ralatively high frequency or ratio. It was difficult to find out the exact relationships among the order of the sound acquisition, phonemic complexity, and the disfluencies. To study the exact influence of the phonemic features upon the disfluencies, it comes important to consider the frequency of the stuttering itself together with the ratio of the disfluencies in which the opportunity of the specific sound's presence was considered. To compare the results of the different studies which has similar purposes, it seems important to consider the tasks and the methodologies in depth.

• 한국컴퓨터정보학회논문지
• /
• 제13권3호
• /
• pp.27-33
• /
• 2008

본 논문에서는 한국어 형태소 분석의 성능향상을 위해서, 어절에서 미등록어를 인식하여 자동으로 사전을 구축하는 방법을 제안한다. 제안하는 사전 구축 방법은 전문 분석 기반 사전 구축 방법과 웹 출현빈도 기반 사전 구축방법으로 구성되어 있다. 전문 분석 기반사전 구축 방법은 전체 문서에서 반복적으로 나타나는 문자열을 미등록어로 인식하고, 웹 출현빈도 기반사전 구축 방법은 반복되지 않은 문자열을 웹 문서에서 검색하여 그 출현빈도를 바탕으로 미등록어를 인식한다. 실험결과 전문 분석만을 바탕으로 하는 기존 접근방법에 비해서 웹 문서에서의 출현빈도도 함께 고려하여 제안하는 사전 구축 방법은 32.39% 정도 재현율이 높게 나타났다.

• 한국응급구조학회지
• /
• 제24권1호
• /
• pp.85-92
• /
• 2020

Purpose: The purpose of this study was to analyze the English abstract key words found within the Korea Society of Emergency Medical Services journal using text mining techniques to determine the adherence of these terms with Medical Subject Headings (MeSH) and identify key word trends. Methods: We analyzed 212 papers that were published from 2012 to 2019. R software, web scraping, and frequency analysis of key words were conducted using R's basic and text mining packages. Additionally, the Word Clouds package was used for visualization. Results: The average number of key words used per study was 3.9. Word cloud visualization revealed that CPR was most prominent in the first half and emergency medical technician was most frequently used during the second half. There were a total of 542 (64.9%) words that exactly matched the MeSH listed words. A total of 293 (35%) key words did not match MeSH listed words. Conclusion: Researchers should obey submission rules. Further, journals should update their respective submission rules. MeSH key words that are frequently cited should be suggested for use.

• 아동학회지
• /
• 제30권2호
• /
• pp.161-177
• /
• 2009

In this study of the relationship between preschoolers' visual perception and reading Hangul words, the 287 participants showed significant developmental change in visual perception between three to five years of age. The researcher developed the computer-based screening Test of Visual Perception for Reading (TVPR). Factor analysis confirmed three factors of TVPR : perception of number, direction, and form constancy. These factors correlated highly with four factors of motor-reduced visual perception of the Korean Developmental Test of Visual Perception (Moon et al. 2003). All factors of TVPR explained reading real words and pseudo words; direction and form constancy perception predicted reading low frequency letters. These findings confirm that preschoolers' skills in visual perception contribute to the reading of words in Hangul.