한·중 분류사구에 관한 비교 연구

오상언(WU XIANGYAN) 원광대학교 인문학연구소 2014 열린정신 인문학연구 Vol.15 No.1

Classifier structure is a very common language phenomenon, which exists widely in various languages. Usually these numeral-classifier compound structures are formed by numeral plus classifier. In this structure, classifier not only defines quantity, but also functions to classify the noun it modifies. Korean and Chinese belong to numeral classifier language. Numeral classifier language has three basic constitution elements, i.e. noun, numeral and classifier. Although both of Korean and Chinese are numeral classifier language, they have different word order, so these three basic elements have similar points and differences on marshalling sequence, which brings enormous inconvenience for Korean and Chinese learners. In order to solve these inconveniences, this article, according to contrastive linguistics theory, uses the correlation method to conduct the analysis and research on numeral-classifier compound structure of these two languages. There are two aspects of major study: 1. To compare marshalling sequence of numeral, classifier and noun in the premier of the existing of these three elements; 2. To compare marshalling sequence of numeral, classifier and verbal in the premier of the existing of these three elements. Through comparative study on these points, it can be found that numeral-classifier compound structure of these two languages has what kind of similar points and differences on meaning, grammatical function and grammatical forms to provide help for language learners.

혼합 약한 분류기를 이용한 AdaBoost 알고리즘의 성능 개선 방법

김정현(Jeong-Hyun Kim),등죽(Teng-Zhu),김진영(Jin-Young Kim),강동중(Dong-Joong Kang) 제어로봇시스템학회 2009 제어·로봇·시스템학회 논문지 Vol.15 No.5

The weak classifier of AdaBoost algorithm is a central classification element that uses a single criterion separating positive and negative learning candidates. Finding the best criterion to separate two feature distributions influences learning capacity of the algorithm. A common way to classify the distributions is to use the mean value of the features. However, positive and negative distributions of Haar-like feature as an image descriptor are hard to classify by a single threshold. The poor classification ability of the single threshold also increases the number of boosting operations, and finally results in a poor classifier. This paper proposes a weak classifier that uses multiple criterions by adding a probabilistic criterion of the positive candidate distribution with the conventional mean classifier: the positive distribution has low variation and the values are closer to the mean while the negative distribution has large variation and values are widely spread. The difference in the variance for the positive and negative distributions is used as an additional criterion. In the learning procedure, we use a new classifier that provides a better classifier between them by selective switching between the mean and standard deviation. We call this new type of combined classifier the “Mixed Weak Classifier”. The proposed weak classifier is more robust than the mean classifier alone and decreases the number of boosting operations to be converged.

What is a Numeral classifier?

Byeong-uk Yi 한국분석철학회 2011 철학적 분석 Vol.0 No.23

A wide variety of languages (e.g., Chinese, Japanese, and Korean) employ special expressions, numeral classifiers, in numeral noun phrases that pertain to the number of some things, e.g., their counterparts of ‘three cows’. This paper discusses what numeral classifiers are, and what distinguishes them from measure words and other related expressions. The paper argues that numeral classifiers are para-numerals for one serving as numeratives (the para-numeral account). Numeratives are expressions belonging to a wide syntactic class that includes not only numeral classifiers but also their syntactic cousins, including (a) various kinds of measure words, and (b) group numeratives. Included among group numeratives are para-numerals, grammatical cousins of numerals: ‘pair’, ‘couple’, ‘dozen’, ‘score’, etc., and their counterparts in numeral classifier languages. On the para-numeral account, numeral classifiers are siblings of the usual para-numeral numeratives. While these are grammatical cousins of numerals for numbers greater than one, numeral classifiers are cousins of numerals for one. The view of classifier languages presented in the paper contrasts sharply with the prevailing view of classifier languages in contemporary linguistics. The prevailing view holds that all classifier language nouns are mass nouns (the mass noun thesis), while taking numeral classifiers to be measure words (the measure word account). The para-numeral account contrasts with the measure word account of the function of classifiers, and meshes well with a thesis of classifier language nouns opposite to the mass noun thesis: classifier languages have count nouns as well as mass nouns (the count noun thesis).

What is a Numeral Classifier?

이병욱 한국분석철학회 2011 철학적 분석 Vol.0 No.23

A wide variety of languages (e.g., Chinese, Japanese, and Korean) employ special expressions, numeral classifiers, in numeral noun phrases that pertain to the number of some things, e.g., their counterparts of ‘three cows’. This paper discusses what numeral classifiers are, and what distinguishes them from measure words and other related expressions. The paper argues that numeral classifiers are para-numerals for one serving as numeratives (the para-numeral account). Numeratives are expressions belonging to a wide syntactic class that includes not only numeral classifiers but also their syntactic cousins, including (a) various kinds of measure words, and (b) group numeratives. Included among group numeratives are para-numerals, grammatical cousins of numerals: ‘pair’, ‘couple’, ‘dozen’, ‘score’, etc., and their counterparts in numeral classifier languages. On the para-numeral account, numeral classifiers are siblings of the usual para-numeral numeratives. While these are grammatical cousins of numerals for numbers greater than one, numeral classifiers are cousins of numerals for one. The view of classifier languages presented in the paper contrasts sharply with the prevailing view of classifier languages in contemporary linguistics. The prevailing view holds that all classifier language nouns are mass nouns (the mass noun thesis), while taking numeral classifiers to be measure words (the measure word account). The para-numeral account contrasts with the measure word account of the function of classifiers, and meshes well with a thesis of classifier language nouns opposite to the mass noun thesis: classifier languages have count nouns as well as mass nouns (the count noun thesis).

A Study of the Process of Noun Categorization by Thai Classifiers Based on a Cognitive Linguistics Approach

Yangwon Hyun 한국태국학회 2012 한국태국학회논총 Vol.19 No.1

The present thesis objective is to study noun categorization and the cognitive process associated with noun categorization by classifiers in Thai. Data was gathered via interviews with 50 Thai undergraduate students. The theory of idealized cognitive models, or ICMs, is employed for data analysis. Result show that Each of classifiers categorizes nouns into class(es) with respect to each ICM(s) of classifier. The system of Thai classifier system in this study consist of 4 types ICMs including classical taxonomy model, simple model, complex model, and radial category model. Human classifier group categorize into various classes according to social and cultural conditions such as social class, respectableness, age, gender, and Buddhism. Classifier /khon/ categorizes nouns as humans with respect to animate taxonomy model. Classifier /thân/categorizes nouns as respectable persons with respect to complex model in which age model, good-respect model, status model, and knowledge-capability model combine to form a cluster model. All these individual models in the cluster characterize the concept of RESPECTABLENESS OF PERSON. Respectively, classifier /nai/ and classifier /nang/ categorize nouns as typical males and females occupations with respect to simple model which is defined an ICM regarding expectations of certain occupations in accordance with gender. Classifier /ong/, beside, categorizes nouns into 2 classes including Buddhist animates which are at the center and Buddhist inanimates which extended from the center with respect to radial category model. As for animal and plant classifier group also categorize nouns into class(es) with respect to radial category model. Their extensions are motivated by resemblance shape with the central members. Animal classifier /tua/ categorizes nouns into animals which are center, furniture and clothes which are extended from center by resemble body-shape with animals. All of the plant classifiers such as /dòk/ /tôn/ /bai/ /mèt/and /lûk/ categorize respectively flowers, trees, leaves, grains, fruits as central members, and they extended to other classes by resemble shape with their central members. It is clear that certain plant parts have salient shape so that they used as a classifier in order to categorize other nouns that have resemble shape.

Learning Distribution Graphs Using a Neuro-Fuzzy Network for Naive Bayesian Classifier

전설위,임준식,Tian, Xue-Wei,Lim, Joon S. The Society of Digital Policy and Management 2013 디지털융복합연구 Vol.11 No.11

Naive Bayesian classifiers 네이브 베이지안 분류기는 샘플 데이터로부터 쉽게 구현될 수 있는 강력하고도 많이 사용되는 형식의 분류기다. 그러나 강한 조건부 독립성으로 인하여 효율이 저하되는 분류 결과를 초래한다. 일반적으로 네이브 베이지안 분류기는 연속성을 가진 특징 데이터의 우도를 처리하기 위해 가우시안 분산을 사용한다. 속성들의 확률밀도는 항상 가우시안 분산에 적합한 것만은 아니다. 또 다른 형식의 분류기는 지도학습을 통해 퍼지 규칙과 퍼지집합을 학습할 수 있는 퍼지신경망이다. 퍼지신경망과 네이브 베이지안 분류기간에는 구조적 유사성을 가지고 있기 때문에 퍼지신경망으로 학습된 분산 그래프를 네이브 베이지안 분류기에 적용하고자 하는 방안이 본 연구의 목적이다. 따라서 네이브 베이지안 분류기에 가우시안 분산 그래프를 사용한 결과와 퍼지 분산 그래프를 사용한 결과를 비교하였다. 이를 위해 leukemia와 colon의 DNA 마이크로어레이 데이터를 적용하여 분류하였다. 네이브 베이지안 분류기에 퍼지 분산 그래프를 사용한 결과 가우시안 분산 그래프를 사용한 결과보다 더 신뢰성이 있음을 보여주었다. Naive Bayesian classifiers are a powerful and well-known type of classifiers that can be easily induced from a dataset of sample cases. However, the strong conditional independence assumptions can sometimes lead to weak classification performance. Normally, naive Bayesian classifiers use Gaussian distributions to handle continuous attributes and to represent the likelihood of the features conditioned on the classes. The probability density of attributes, however, is not always well fitted by a Gaussian distribution. Another eminent type of classifier is the neuro-fuzzy classifier, which can learn fuzzy rules and fuzzy sets using supervised learning. Since there are specific structural similarities between a neuro-fuzzy classifier and a naive Bayesian classifier, the purpose of this study is to apply learning distribution graphs constructed by a neuro-fuzzy network to naive Bayesian classifiers. We compare the Gaussian distribution graphs with the fuzzy distribution graphs for the naive Bayesian classifier. We applied these two types of distribution graphs to classify leukemia and colon DNA microarray data sets. The results demonstrate that a naive Bayesian classifier with fuzzy distribution graphs is more reliable than that with Gaussian distribution graphs.

어휘확산 이론을 통해 본 양사 ‘개(個)’의 사용 변화 ― 관화(官話)의 동물명사를 중심으로

이지은 ( Lee Jieun ) 한국중국학회 2018 중국학보 Vol.85 No.-

중국어의 개체양사가 어떠한 기능을 담당하느냐에 관해 학자들의 견해는 크게 부류화, 개체화, 셈의 단위 표시 등으로 나뉘고 있다. 본고는 중국어 개체양사의 기능은 ‘셈의 단위’ 표시이며, 명사와의 의미적 관계에서 나타나는 부류화 기능은 개체양사의 기원에 따른 부차적 기능일 뿐 본질적 기능은 아니라고 본다(李知恩2011a, 2011b). 언어의 경제 원칙에 따라 부차적 기능인 분류 기능은 사실상 반드시 발휘되어야 하는 것은 아니지만, 셈의 단위 표시 기능은 개체양사의 본질적 기능이므로 어떠한 경우에도 의무적으로 발휘되어야 한다. 그렇기 때문에 수 분류사가 발달한 언어에는 많은 부류의 명사와 결합할 수 있는 보편 분류사(general classifier)가 대부분 존재하며(Aikhenvald, 2000:98) 중국어도 마찬가지로 보편양사 ‘個’가 고빈도로 사용되는 것이다.따라서 본고는 현대중국어에서 양사의 본질적 기능만을 담당하는 무표적 양사인 ‘個’의 사용 영역은 앞으로 계속해서 확대될 것이라고 본다. 이러한 문제를 새로이 논의해 보기 위해 본고는 두 가지 다른 시각에서 연구를 진행하였다. 첫째, ‘個’와 명사의 결합 범위가 상이한 방언별 고찰을 통하여 ‘個’의 사용 영역이 확대되는 현상을 확인한다. 둘째, ‘個’의 사용 범위 확대가 어휘확산이라는 특징을 보이는 것에 주목하고, 어휘확산 이론(lexical diffusion)을 적용하여 방언별 사용 양상을 부류가 아닌 단어를 중심으로 살펴본다.이를 위해 본 연구는 ‘個’가 보편양사로 사용되고 있는 관화 지역의 방언지점 9곳을 선정하고, 해당 방언에서 동물명사와 결합하는 양사를 고찰하였다. 이를 통하여 ‘個’의 쓰임이 확대되는 패턴과 그 기제를 확인하고, 과거 연구에서 예외로 간주되었던 현상들이 어휘확산 과정 중에 나타나는 필연적인 현상임을 밝혀본다. Numeral classifiers in languages have generic classifiers that are ambiguous and can be used with a variety of nouns. They can be used in nouns to replace other more specific classifiers. This is the universal commonality of almost all the uses of the word classification (Aikhenvald, 2000:98). Chinese also has a wide-ranging general classifier “ge”. It originally refers to a specific object, which is later used as an abstract unit of measurement. This paper proposes that the main function of individual classifiers is to define the unit of measurement, so from an economic perspective the classification function of individual classifiers is secondary. This is why numeral classifiers in languages employ a generalized classifier with a higher degree of ambiguity. That is to say, there are large semantic differences between nouns that are combined with general classifiers. Past research indicates that there are always exceptions to the use of the semantic features of nouns which illustrate the choice of individual classifiers and nouns.Other Chinese dialects also have the classifier “ge.” However, its scope of application is different than in Putonghua. In some dialects (especially in the southern dialects), the scope of application of “ge” is narrower than in Putonghua. Similarly, in other dialects of Chinese, the scope of application of “ge” can vary substantially. This paper will contend that these differences occur because the extent to which “ge” is applied to nouns varies between dialects. Therefore, the combination of “ge” and various nouns in different dialects will be examined. This research suggests that the core nouns that are commonly associated with “ge” can also be seen in the “ge” choices employed in different dialects.In order to observe this phenomenon, this article will use the Chinese dialect dictionary, other scholars’ survey materials and a small part of the author’s survey to conduct a comprehensive survey of the use of “ge” in nine Mandarin dialects. Through the examination and comparison of cross-dialects, this paper will discuss a number of issues. Firstly, what are the core nouns combined with “ge?” Secondly, what are the boundary nouns that differ from the “ge” combinations in different dialects? This article will investigate the spread of the classifier “ge” in different Mandarin dialects in order to find out the difference in the degree of diffusion of the “ge” function. It should be pointed out that “ge” does not necessarily spread to all the nouns of a certain semantic class. Thus, the unit of investigation in this paper is a single noun, not a type of noun of a semantic class. Through such an investigation, the path of “changing” and the direction of further generalization in the future can be seen.

Research on Dynamic Cost-Sensitive SVM Classifier based on Chaos Particle Swarm Optimization Algorithm

Ruili Zhang 보안공학연구지원센터 2014 International Journal of Control and Automation Vol.7 No.10

In order to improve the performance of Support Vector Machine (SVM) classifier for imbalanced data, this paper proposes dynamic cost-sensitive SVM classifier based on chaos particle swarm optimization (CPDC_SVM). Firstly, this paper introduces dynamic cost-sensitive thought to SVM classifier, and gives the method for structuring dynamic cost and cost-sensitive SVM model. Secondly, we propose the evaluation methodology performance for classifier, and adopts decimal base to code the particles. At last, chaos thought is introduced in particle swarm optimization algorithm, and the Algorithm of the dynamic cost-sensitive SVM classifier is given, which improves convergent speed and accuracy of particle swarm optimization, and can optimize dynamic cost-sensitive SVM well, so CPDC_SVM adds effectively the convergence speed and accuracy for the particle swarm optimization algorithm. Experimental results show CPDC_SVM has higher precision than traditional SVM classifier, and dynamic cost and chaos particle swarm optimization can improve the performance for classifier.

언간에 나타난 분류사의 분포와 의미 연구

배영환(Younghwan Bae) 한국중원언어학회 2015 언어학연구 Vol.0 No.36

The Study on Distribution and Meaning of Classifier of Eon’gan (Vernacular Letters). This study was designed to identify a list of diverse classifiers shown in Eon’gan materials in Joseon Dynasty, and to discuss their distribution and meanings. The structure of the quantifier phrase in Eon’gan is that a noun is located at before a numeral or classifier like ‘a noun + a numeral’ or ‘a noun + a numeral + a classifier.’ To name a few, interesting characteristics of some classifiers are as follows. First, classifiers related to clothing habits are mensural such as ‘pil’ and sortal such as ‘suit." Second, classifiers relevant to the dietary life seemingly originated from names of containers. Third, classifiers related to numbers as well as to fish and shellfish are so various: There is something different between numbers in Eon’gan and numbers in a modern language. Fourth, many classifiers related to vegetable and seaweed show collectivity such as ‘bundle,’ while most classifiers related to grain are measuring classifiers, showing volume. Finally, classifiers related to disease and medicine are from Sino-Korean words, whereas most classifiers previously mentioned are native. Some other interesting findings are discussed in detail.

Two types of classifier reduplications in Mandarin

Yanyang Zheng,김규민 경희대학교 언어정보연구소 2022 언어연구 Vol.39 No.1

Reduplication is a word formation process, which has been widely attested across languages. Mandarin is a language that shows extensive range of reduplication across different categories such as a classifier. However, despite its widespread use of classifier reduplication, the syntax of reduplicated classifiers has not been well understood in the current literature. This paper addresses the syntax of reduplicated classifiers by focusing on the two issues: (i) the category of a reduplicated classifier, and (ii) the syntactic derivation of classifier reduplication. As for issue (i), we propose that classifier reduplication creates two different syntactic categories, namely D(eterminer) and A(djective), building on novel empirical evidence not provided in the previous studies. Regarding issue (ii), building on head movement analysis of Travis (2003) on reduplication, we propose that a CL head undergoes head movement which results in the proposed categorial status of reduplication, i.e., D and A. This paper not only contributes to the understanding of syntax of classifier reduplication in Mandarin, but also provides support for the current view that reduplication can be motivated by syntax (Basciano and Melloni 2017; Kimper 2008; Travis 2003).