• The Journal of the Korea Contents Association
• /
• v.10 no.3
• /
• pp.10-20
• /
• 2010

This paper presents chunking strategy of a contiguous nouns sequence using semantic class. We call contiguous nouns which can be treated like a noun the compound noun phrase. We use noun pairs extracted from a syntactic tagged corpus and their semantic class pairs for chunking of the compound noun phrase. For reliability, these noun pairs and semantic classes are built from a syntactic tagged corpus and detailed dictionary in the Sejong corpus. The compound noun phrase of arbitrary length can also be chunked by these information. The 38,940 pairs of 'left noun - right noun', 65,629 pairs of 'left noun - semantic class of right noun', 46,094 pairs of 'semantic class of left noun - right noun', and 45,243 pairs of 'semantic class of left noun - semantic class of right noun' are used for compound noun phrase chunking. The test data are untrained 1,000 sentences with contiguous nouns of length more than 2randomly selected from Sejong morphological tagged corpus. Our experimental result is 86.89% precision, 80.48% recall, and 83.56% f-measure.

• Journal of Digital Convergence
• /
• v.17 no.2
• /
• pp.283-292
• /
• 2019

The loss of semantic knowledge and impairments in semantic associations by semantic category is gaining increasing attention, as indicators of early-stage cognitive decline. As such, we assigned semantic association task (SAT) to normal elderly (NE) and those with subjective memory impairment (SMI) or mild cognitive impairment (MCI) to examine their performance by semantic subcategories and the differences in error patterns. We found a significant difference in the number of correct response and reaction time under the SAT categories among the three groups, with the highest performance observed in 'function' and the lowest performance in 'superordinate' and 'part/whole'. Moreover, the error frequency was the lowest in NE, followed by those with SMI and MCI, with the latter two groups showing a significant increase in no-response. Our findings demonstrate the varying extent and process of impairments in the semantic network by category over different stages of cognitive decline. Thus, we proposed SAT performance as an indicator to detect and follow-up on cognitive decline in elderly with cognitive disorder.

• Proceedings of the Korean Society for Language and Information Conference
• /
• 2007.11a
• /
• pp.321-329
• /
• 2007

This paper reports on the application of network analysis approaches to investigate the characteristics of graph representations of Japanese word associations. Two semantic networks are constructed from two separate Japanese word association databases. The basic statistical features of the networks indicate that they have scale-free and small-world properties and that they exhibit hierarchical organization. A graph clustering method is also applied to the networks with the objective of generating hierarchical structures within the semantic networks. The method is shown to be an efficient tool for analyzing large-scale structures within corpora. As a utilization of the network clustering results, we briefly introduce two web-based applications: the first is a search system that highlights various possible relations between words according to association type, while the second is to present the hierarchical architecture of a semantic network. The systems realize dynamic representations of network structures based on the relationships between words and concepts.

• The Journal of Korean Association of Computer Education
• /
• v.9 no.3
• /
• pp.121-128
• /
• 2006

This study suggested the new e-learning systems based on agent to provide an adaptable learning. In Semantic Web environment, to develop an ontology and an intelligent agent is essential for an adaptable e-learning systems. Especially, to develop a reasoning engine using analysis of learning content and learners' information can offer an effective e-learning system. Therefore, we developed an applying model to an adaptable e-learning systems and the various ontologies for Semantic Web environment. Moreover, we analyzed and developed ontologies within the framework of learning domain, a learner and interface. Further, we implemented an intelligent e-learning for applying an agent's reasoning. Through this system proposed, we suggested the new e-learning systems model for Semantic Web environment.

• The Journal of the Korea Contents Association
• /
• v.15 no.10
• /
• pp.636-646
• /
• 2015

In information society, the need to detect document duplication and plagiarism is increasing. Many studies have progressed to meet such need, but there are limitations in increasing document duplication detection quality due to technological problem of natural language processing. Recently, some studies tried to increase the quality by applying syntatic-semantic analysis technique. But, the studies have the problem comparing syntactic-semantic trees. This paper develops a syntactic-semantic tree comparator, designs and implements a discriminator of similar documents using the comparator. To evaluate the system, we analyze the correlation between human discrimination and system discrimination with the comparator. This analysis shows that the proposed discrimination has good performance. We need to define the document type and improve the processing technique appropriate for each type.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.16 no.11
• /
• pp.2374-2381
• /
• 2012

This paper proposes a new enhancing snippet extraction method using fuzzy and semantic features. The proposed method creates a delegate of sentence by using semantic features. It extracts snippet using fuzzy association between a delegate sentence and sentence set which well represents query. In addition, the method uses pseudo relevance feedback to expand query which extracts snippet to be well reflected semantic user's intention. The experimental results demonstrate the proposed method can achieve better snippet extraction performance than the previous methods.

• The Journal of Information Systems
• /
• v.13 no.2
• /
• pp.135-153
• /
• 2004

The formation of a supply chain should be approached with distributive viewpoint considering the dynamic nature of ordering environments. The purpose of this paper is to design a semantic web to support agent negotiation forming supply chains. With the view of a supply chain as composed of competitive agents which represent each member of supply chains with a local goal, a supply chain is formed through negotiation of price, due date choosing partners to contract. In the development of a negotiation method, we employ ontology and rules which are the basic techniques of the semantic web for supporting automated negotiations. Especially, a framework for a negotiation process is suggested and domain, processes and rules ontology are designed interrelatedly. With this modeling, a possibility of the semantic web based agent negotiation is suggested.

• Korean Journal of English Language and Linguistics
• /
• v.3 no.4
• /
• pp.589-609
• /
• 2003

The purpose of this paper is to show that lexical equivalency in translation can be achieved when it is based on semantic prosodies of lexical items. This paper examines the semantic prosodies of two seemingly synonymous English adjectives ‘empty’ and ‘blank’ on the basis of the corpus given in Cobuild English Collocations on CD-ROM and proposes that they are different in terms of spatial dimensions. Thus when a Korean equivalent pin derived from the verb pita is translated into English, syntagmatic phraseological environments of the Korean adjective must be taken into account to attain the equivalency of the source and target languages. Relevant Korean corpus was taken from the 21st Century Sejong Plan (2002). Out of 12 examples of pin konggan, five appear to be equivalent to ‘blank’ and seven to ‘empty.’ The five to seven ratio in different usage indicates that the equivalency problem concerning the lexical item pin is not a trivial matter in translation.

• Korean Journal of Child Studies
• /
• v.18 no.2
• /
• pp.73-85
• /
• 1997

The purpose of the present study was to examine the role of semantic and syntactic knowledge in the first language acquisition of Korean classifiers. The elicited classifiers production test(EPT) was conducted to 105 children aged from 2 to 7. EPT consisted of 16 classifiers and two items for each classifier. 32 items were divided into 2 major semantic features: animacy and inanimacy. The semantic features of inanimacy were subcategorized into 3 features such as neutral, shape and function. The results revealed that; 1) children produced the correct structure of classification from the very early age with correct word order of the noun phrase showing early fundamental syntactic knowledge; 2) The earliest response pattern was to respond to all nouns in the same way using a neutral classifier showing no apparent semantic basis for their choice; 3) Children didn't show any preference for animate, shape, or function classifiers.

• The Journal of Information Systems
• /
• v.31 no.4
• /
• pp.01-23
• /
• 2022

Purpose This study aims to propose a deep learning-based fake review detection model by combining authors' linguistic features and semantic information of reviews. Design/methodology/approach This study used 358,071 review data of Yelp to develop fake review detection model. We employed linguistic inquiry and word count (LIWC) to extract 24 linguistic features of authors. Then we used deep learning architectures such as multilayer perceptron(MLP), long short-term memory(LSTM) and transformer to learn linguistic features and semantic features for fake review detection. Findings The results of our study show that detection models using both linguistic and semantic features outperformed other models using single type of features. In addition, this study confirmed that differences in linguistic features between fake reviewer and authentic reviewer are significant. That is, we found that linguistic features complement semantic information of reviews and further enhance predictive power of fake detection model.