• Journal of KIISE
• /
• v.44 no.12
• /
• pp.1290-1295
• /
• 2017

Document classification, a task of classifying the category of each document based on text, is one of the fundamental areas for natural language processing. Document classification may be used in various fields such as topic classification and sentiment classification. Neural network models for document classification can be divided into two categories: word-level models and character-level models that treat words and characters as basic units respectively. In this study, we propose a neural network model that combines character-level and word-level models to improve performance of document classification. The proposed model extracts the feature vector of each word by combining information obtained from a word embedding matrix and information encoded by a character-level neural network. Based on feature vectors of words, the model classifies documents with a hierarchical structure wherein recurrent neural networks with attention mechanisms are used for both the word and the sentence levels. Experiments on real life datasets demonstrate effectiveness of our proposed model.

• Korean Journal of Ecology and Environment
• /
• v.46 no.4
• /
• pp.471-487
• /
• 2013

This review paper was written to provide background information as well as future application for aquatic ecologists interested in using stable isotope. Stable isotope techniques has proved to be an extremely useful to elucidate a lot of environmental and ecological problems. Stable isotopes have been used as possible tracers to identify sources, to quantify relative inputs in a system. When utilized carefully, stable isotope tools provides apparent advantages for the scientists to find out the processes of material cycles in various environments and energy flows in natural ecosystems.

• Journal of Haehwa Medicine
• /
• v.19 no.2
• /
• pp.145-151
• /
• 2011

Backgrounds: Multidisciplinary approaches including surgery, chemotherapy, and radiation therapy are currently being performed to target various cancers in Western Medicine. However, some cancers still remain difficult to battle, which has long attracted many scientists for the discovery of new agents to fight cancers. Ginseng is one of the herbs used in Oriental Medicine including Korea, China and Japan. We have further investigated ginseng for its anticancer effect. Objective: This is a comprehensive review summary of anticancer effect of ginseng and ginsenoids as a possible agent for future cancer treatment. Methods: Data were retrieved from two web sites; www.pubmed.com and www.riss.kr, and authorized texts concerning anticancer effects of ginseng. From collected data, information on anticancer effect of ginseng was thoroughly sorted, restructured, then assessed. Results: Panax Ginseng C.A. Meyer belongs to Araliaceae Panax family, a perennial prairie plant with its root known as Ginseng Radix. Ginseng induces anticancer effect through cell cycle arrest, acceleration of apoptosis, anti-angiogenesis, and suppression of metastasis. Anticancer effect of ginseng may be due to single compound or multi-compound actions. Many studies report involvement of immune mechanisms of cytokines, Natural Killer (NK) cells, macrophages and some antibodies in enhancing anticancer effect of ginseng. In near future, possibility of applying these mechanisms into clinical trials is convinced. There were some important findings on saponin in ginsenoids in reviewing for this article; First, eradication of metastatic tumors were influenced by macrophage activation. Second, suppression of malignant melanoma cell metastasis to lung were induced by macrophage and NK cell activation in spleen with red ginseng acidic polysaccharide (RGAP). Third, final metabolites of M1, M4 had exerted anticancer effect of ginseng. Conclusion: Unknown anticancer mechanisms of ginseng have been studied for many years up until now. Ginseng is comprised of multiple bio-chemical compounds that create complex pharmaceutical interactions. Therefore, for its proper usage and safe prescription, studies on different types of ginseng and patients' susceptibility to ginseng according to their constitution and stages of the disease should be further pursued. More efforts are needed to understand the anticancer mechanisms of ginseng as well.

• KIISE Transactions on Computing Practices
• /
• v.20 no.10
• /
• pp.549-554
• /
• 2014

Handwritten character recognition data is usually imbalanced because it is collected from the natural language sentences written by different writers. The imbalanced data can cause seriously negative effect on the performance of most of machine learning algorithms. But this problem is typically ignored in handwritten character recognition, because it is considered that most of difficulties in handwritten character recognition is caused by the high variance in data set and similar shapes between characters. We propose the oversampling-based ensemble learning methods to solve imbalanced data problem in handwritten character recognition and to improve the recognition accuracy. Also we show that proposed method achieved improvements in recognition accuracy of minor classes as well as overall recognition accuracy empirically.

• Journal of KIISE:Software and Applications
• /
• v.37 no.12
• /
• pp.928-935
• /
• 2010

Identifying the referents of omitted elements in a text is an important task to many natural language processing applications such as machine translation, information extraction and so on. These omitted elements are often called zero anaphors or zero pronouns, and are regarded as one of the most common forms of reference. However, since all zero elements do not refer to explicit objects which occur in the same text, recent work on zero anaphora resolution have attempted to identify the anaphoricity of zero pronouns. This paper focuses on intra-sentential anaphoricity determination of subject zero pronouns that frequently occur in Korean. Unlike previous studies on pair-wise comparisons, this study attempts to determine the intra-sentential anaphoricity of zero pronouns by learning directly the structure of clauses in which either non-anaphoric or inter-sentential subject zero pronouns occur. The proposed method outperforms baseline methods, and anaphoricity determination of zero pronouns will play an important role in resolving zero anaphora.

• Journal of KIISE:Software and Applications
• /
• v.35 no.5
• /
• pp.306-312
• /
• 2008

In a goal-oriented dialogue, speaker's intention can be approximated by a semantic structure that consists of a pair of a speech act and a concept sequence. Therefore, it is very important to correctly identify the semantic structure of an utterance for implementing an intelligent dialogue system. In this paper, we propose a model to efficiently analyze the semantic structures based on an active teaming method. To reduce the burdens of high-level linguistic analysis, the proposed model only uses morphological features and previous semantic structures as input features. To improve the precisions of semantic structure analysis, the proposed model adopts CRFs(Conditional Random Fields), which show high performances in natural language processing, as an underlying statistical model. In the experiments in a schedule arrangement domain, we found that the proposed model shows similar performances(92.4% in speech act analysis and 89.8% in concept sequence analysis) to the previous models although it uses about a third of training data.

• Journal of KIISE:Software and Applications
• /
• v.34 no.8
• /
• pp.721-730
• /
• 2007

UML (Unified Modeling Language) is widely used to analyze and design target software. Developers also implement the target software based on the UML artifacts. However, it is difficult to validate whether the artifacts are generated to correspond to the modeling guidelines because the guidelines for UML modeling are described in natural language. This paper discusses rule based model checker focused on whether models are designed according to modeling methodology. We propose rules and their own checker, named GARDIAN, for UML model validation. The checkers are designed for COMET method for the real-time embedded system. We illustrate our checkers using Intelligent Robot system to validate our approach.

• Journal of KIISE
• /
• v.44 no.11
• /
• pp.1138-1148
• /
• 2017

Topic extraction is a technology that automatically extracts a set of topics from a set of documents, and this has been a major research topic in the area of natural language processing. Representative topic extraction methods include Latent Dirichlet Allocation (LDA) and word clustering-based methods. However, there are problems with these methods, such as repeated topics and mixed topics. The problem of repeated topics is one in which a specific topic is extracted as several topics, while the problem of mixed topic is one in which several topics are mixed in a single extracted topic. To solve these problems, this study proposes a method to extract topics using an LDA that is robust against the problem of repeated topic, going through the steps of separating and merging the topics using the similarity between words to correct the extracted topics. As a result of the experiment, the proposed method showed better performance than the conventional LDA method.

• The Journal of the Korea Contents Association
• /
• v.11 no.11
• /
• pp.503-513
• /
• 2011

Appreciation of nature and an understanding of the biological sciences by the general public are key to the popularization of modern science. In particular, informal and accessible venues such as museum exhibits occupy a crucial role in science education, and they depend heavily on fields related to macrobiology, including Ecology, Animal Behavior, and Environmental Science. Unfortunately, lack of engaged experts and superficial descriptions of natural phenomena all too often prevent scientific knowledge from being shared effectively with the general public. Raw information itself and knowledge are not in a form or structure accessible to nonspecialists. In order to move successfully deliver substantive comprehension of the biological knowledge to the general public, it is necessary to categorize information from a content-conscious perspective and transform it into useful biological content. Therefore, the role of scientists is critically important in a series of processes that include theme selection, editing, and even graphical layout of contents. These processes require not only a scientific and logical way of thinking, but also an aptitude for artistic presentation and effective communication. The concept of Translation is presented as a theoretical and operational framework for the popularization of science.

• Journal of Microbiology and Biotechnology
• /
• v.17 no.4
• /
• pp.547-559
• /
• 2007

Bacillus thuringiensis (Bt) was first described by Berliner [10] when he isolated a Bacillus species from the Mediterranean flour moth, Anagasta kuehniella, and named it after the province Thuringia in Germany where the infected moth was found. Although this was the first description under the name B. thuringiensis, it was not the first isolation. In 1901, a Japanese biologist, Ishiwata Shigetane, discovered a previously undescribed bacterium as the causative agent of a disease afflicting silkworms. Bt was originally considered a risk for silkworm rearing but it has become the heart of microbial insect control. The earliest commercial production began in France in 1938, under the name Sporeine [72]. A resurgence of interest in Bt has been attributed to Edward Steinhaus [105], who obtained a culture in 1942 and attracted attention to the potential of Bt through his subsequent studies. In 1956, T. Angus [3] demonstrated that the crystalline protein inclusions formed in the course of sporulation were responsible for the insecticidal action of Bt. By the early 1980's, Gonzalez et al. [48] revealed that the genes coding for crystal proteins were localized on transmissible plasmids, using a plasmid curing technique, and Schnepf and Whiteley [103] first cloned and characterized the genes coding for crystal proteins that had toxicity to larvae of the tobacco hornworm, from plasmid DNA of Bt subsp. kurstaki HD-1. This first cloning was followed quickly by the cloning of many other cry genes and eventually led to the development of Bt transgenic plants. In the 1980s, several scientists successively demonstrated that plants can be genetically engineered, and finally, Bt cotton reached the market in 1996 [104].