• Journal of Korea Society of Industrial Information Systems
• /
• v.9 no.4
• /
• pp.1-7
• /
• 2004

In this thesis, we introduce a parsing method which use information of the post position in Korean to get the exact parsing tree. In order to implement this method we classified categories of the predicates, and defined sentence-types based on these categories. We tried to make parsing using the method grasping the grammatical role of the noun phrase that have to exist in each sentence-type. In parser control mechanism, we use some heuristics based on linguistic frame. We use conditional unification to implement analysis. It is Possible to reduce ambiguous because the parsing method suggested helps to Prune the branches which are unnecessary.

• Journal of KIISE
• /
• v.44 no.1
• /
• pp.36-43
• /
• 2017

Syntactic information represents the dependency relation between predicates and arguments, and it is helpful for improving the performance of Semantic Role Labeling systems. However, syntax analysis can cause computational overhead and inherit incorrect syntactic information. To solve this problem, we exclude syntactic information and use only morpheme information to construct Semantic Role Labeling systems. In this study, we propose an end-to-end SRL system that only uses morpheme information with Stacked Bidirectional LSTM-CRFs model by extending the LSTM RNN that is suitable for sequence labeling problem. Our experimental results show that our proposed model has better performance, as compare to other models.

• Journal of KIISE
• /
• v.42 no.2
• /
• pp.220-226
• /
• 2015

Semantic role labeling (SRL) systems determine the semantic role labels of the arguments of predicates in natural language text. An SRL system usually needs to perform four tasks in sequence: Predicate Identification (PI), Predicate Classification (PC), Argument Identification (AI), and Argument Classification (AC). In this paper, we use the Korean Propbank to develop our Korean semantic role labeling system. We describe our Korean semantic role labeling system that uses sequence labeling with structured Support Vector Machine (SVM). The results of our experiments on the Korean Propbank dataset reveal that our method obtains a 97.13% F1 score on Predicate Identification and Classification (PIC), and a 76.96% F1 score on Argument Identification and Classification (AIC).

• Journal of KIISE:Software and Applications
• /
• v.36 no.4
• /
• pp.320-328
• /
• 2009

In general, there are two types of noun phrases(NP): Base Noun Phrase(BNP), and Maximal-Length Noun Phrase(MNP). MNP identification can largely reduce the complexity of full parsing, help analyze the general structure of complex sentences, and provide important clues for detecting main predicates in Chinese sentences. In this paper, we propose a 2-phase hybrid approach for MNP identification which adopts salient features such as expanded chunks and classified punctuations to improve performance. Experimental result shows a high quality performance of 89.66% in $F_1$-measure.

• Annual Conference on Human and Language Technology
• /
• 1996.10a
• /
• pp.329-337
• /
• 1996

일-한 기계 번역을 연구하는 많은 연구자들은 양국어의 문절-어절 단위의 어순 일치와 같은 구조적 유사성을 최대한 이용하기 위해 직접 번역 방식을 채택하고 있다. 그러나, 일본어와 한국어 술부간에는 대응하는 품사의 불일치 및 국부적인 어순의 불일치 등이 어려운 문제로 남아 있다. 본 논문에서는 이들 술부 표현의 불일치를 해결하기 위해 이미 제안하였던 "양상 테이블을 기반으로 한 한국어 술부의 생성 방법"에 대해 좀더 체계적인 평가를 하고자 한다. 이 방법은 술부만을 대상으로 하는 추상적이고 의미 기호적인 양상 자질(modality feature)을 테이블화(양상 테이블)하여, 양국어의 술부 표현의 피봇(pivot)으로 이용함으로써 술부 양상 표현의 효과적인 번역을 가능하게 하였다. 일본어 499 문을 대상으로 실제 술부의 번역처리를 시행해 본 결과, 약 97.7%가 자연스럽게 번역됨을 확인하였다. 특히, 술부의 생성 부분은 일본어에 의존하지 않는 양상 테이블을 도입함으로써 일-한뿐만 아니라 다른 언어로부터의 한국어 술부 생성에도 적용시킬 수 있을 것이다.

• Phonetics and Speech Sciences
• /
• v.6 no.4
• /
• pp.171-179
• /
• 2014

Purposes: Speech acceptability means the subjective judgement of listeners regarding the naturalness and normality of the speech. The purpose of this study was to determine the predicting variables for speech acceptabilities of children with cochlear implants. Methods: Twenty seven children with CI participated. They had profound pre-lingual hearing loss without any additional disabilities. The mean of chronological ages was 8;9, and mean of age of implantation was 2;11. Speech samples of reading and spontaneous speech were recorded separately. Twenty college students who were not familiar to the speech of deaf children evaluated the speech acceptabilities using visual analog scale. 1 segmental (articulation) and 6 suprasegmental features (pitch, loudness, quality, resonance, intonation, and speaking rate) of speech were perceptually evaluated by 3 SLPs. Correlation and multiple regression analysis were performed to evaluate the predicting variables. Results: The means of speech acceptability for reading and spontaneous speech were 73.47 and 71.96, respectively. Speech acceptability of reading was predicated by the severity of intonation and articulation. Speech acceptability of spontaneous speech was predicated by the severity of intonation and loudness. Discussion and conclusion: Severity of intonation was the most effective variable to predict the speech acceptabilities of both reading and spontaneous speech. A further study would be necessary to generalize the result and to apply this result to intervention in clinical settings.

• Journal of Korean Medical classics
• /
• v.24 no.6
• /
• pp.111-124
• /
• 2011

The objectives of this study are to look into the grammatical characteristics and find misinterpretations on the translation books. 1. Sentences characteristics 1) Lots of ellipses of grammatical parts can be found such as conjunction, postposition, particle, Coverb, and focus on the parts which has practical meaning such as noun, pronoun, verb, adjective in the sentences. 2) Some predicates are skipped in the later phrases which has contradictive concepts against them of former phrases. 3) Pure Korean word order is exposed especially in complement. 2. Translation fallacy 1) There is fallacy in the sentences omitted paratactic conjunction as follows (1) mistranslation based on the wrong concept of the context between equal relation and subordinate relation. (2) failure on setting up the period, (3) misunderstanding equal relation as cause relation. 2) Some singular phrases, which are condition relation, were analyzed as plural phrases in the sentences omitted connection conjunction. 3) Ellipses of postposition obstruct understanding the difference between modifier and modificand in some sentences. 4) Some cause relation phrases were translated as equality relation due to lack of recognition of ellipsis of coverbs.

• Journal of KIISE:Computing Practices and Letters
• /
• v.13 no.6
• /
• pp.402-406
• /
• 2007

Rich Internet Application(RIA) is one of the Web 2.0 technologies and is expected to be a next generation user interface technique which allows flexible and dynamic manipulation for Web searches. This paper addresses a personalization mechanism for advanced Web search using RIA for abundant user interactions. We devised a dynamic and graphical user interface instead of previous text-based searches and a client side application for storing personal preference information. In this research, we implemented the graphical personalized search manager using Yahoo web search API and widget, and demonstrated its effectiveness by performing some experiments with various query terms and representative predicates.

• Korean Journal of Logic
• /
• v.17 no.1
• /
• pp.71-109
• /
• 2014

Gila Sher believes that Tarskian definition of logical consequence is a conceptually and extensionally adequate explanation. She has tried to show this on the basis of Mostowskian conceptions of generalized quantifiers as being invariant under isomorphic structures and her own conceptions of models. In this paper I try to show that her attempt to justify the Tarskian definition is only partially successful. I admit that her conceptions of the logical as being invariant under isomorphic structures are enough to show the logical formality of logical consequence relations. But I think that since her conceptions of meanings of terms are quite inadequate for dealing with the problem of empty predicates, she fails to distinguish logically necessary truths from other kinds of truths.

• Journal of the Korean Society of Earth Science Education
• /
• v.5 no.3
• /
• pp.276-286
• /
• 2012

This research is aimed to examine the differences through comparative analysis of the Earth Science contents in the science textbooks between Korea and DPRK. The contents of level and scope in science textbooks between Korea and DPRK are analysed by TIMSS frameworks. The results of this research are as follows : 1. The science textbooks of DPRK is lower in quality of paper, printing to that of Korea, and the illustrations, editing design in the textbooks of DPRK are fewer, monochromic and monotonous while those in Korea. 2. The contents of Earth Science in DPRK's science textbooks rank 37.0%, but those of Korea's science textbooks rank 25.5% of the whole textbooks. The learning units related to Earth Science are generally similar to the level and scope in science textbooks between Korea and DPRK. The type of inquire activities in the textbooks of DPRK largely takes on the model experiment, and it was shown that the number of experiments directly made by children is very small compared to Korea' textbooks. 3. There are lots of differences in Earth Science learning terms and predicates used in the textbooks between Korea and DPRK.