• Journal of information and communication convergence engineering
• /
• v.5 no.3
• /
• pp.233-238
• /
• 2007

An approach to identify Chinese event types is proposed in this paper which combines a good feature selection policy and a Maximum Entropy (ME) model. The approach not only effectively alleviates the problem that classifier performs poorly on the small and difficult types, but improve overall performance. Experiments on the ACE2005 corpus show that performance is satisfying with the 83.5% macro - average F measure. The main characters and ideas of the approach are: (1) Optimal feature set is built for each type according to local feature selection, which fully ensures the performance of each type. (2) Positive and negative features are explicitly discriminated and combined by using one - sided metrics, which makes use of both features' advantages. (3) Wrapper methods are used to search new features and evaluate the various feature subsets to obtain the optimal feature subset.

• The Journal of Information Systems
• /
• v.19 no.4
• /
• pp.253-270
• /
• 2010

The adjustment of negative emotions in later life is related to the quality of life and life satisfaction. any studies have examined how to control the negative emotions in related to physical, psychological, and social relationships. As the digital media and contents positively influence elder's mental and somatic well-being, it is significant to examine this problem from IT usage, especially the selection of digital contents. This study validated the theoretical study through a qualitative exploratory study comprising both negative emotions and the selection of digital contents, and empirically tested the proposed research model on the older people. The result shows the pattern of linkages between the grief of loss and the pursuit of digital contents. Loss from being parted by death and loss of relations was positively related to the information seeking contents. Economical loss and loss of physical functions was positively related to the innovation seeking contents. Loss of physical functions and Loss from being parted by death was positively related to the emotion seeking contents. And economic loss and loss of relations was positively related to the entertainment seeking contents.

• The Korean Journal of Applied Statistics
• /
• v.16 no.2
• /
• pp.335-349
• /
• 2003

In this paper, we use Bayesian method for model selection of poisson vs. negative binomial distribution, and normal, double exponential vs. cauchy distribution. The fractional Bayes factor of O'Hagan (1995) was applied to Bayesian model selection under the assumption of noninformative improper priors for all parameters in the models. Through the analyses of real data and simulation data, we examine the usefulness of the fractional Bayes factor in comparison with intrinsic Bayes factors of Berger and Pericchi (1996, 1998).

• Journal of the Korean Institute of Intelligent Systems
• /
• v.12 no.5
• /
• pp.411-416
• /
• 2002

Recently, the trial and success of malicious cyber attacks has been increased rapidly with spreading of Internet and the activation of a internet shopping mall and the supply of an online internet, so it is expected to make a problem more and more. Currently, the general security system based on Internet couldn't cope with the attack properly, if ever, other regular systems have depended on common softwares to cope with the attack. In this paper, we propose the positive selection mechanism and negative selection mechanism of T-cell, which is the biological distributed autonomous system, to develop the self/non-self recognition algorithm, the anomalous behavior detection algorithm, and AIS (Artificial Immune System) that is easy to be concrete on the artificial system. The proposed algorithm can cope with new intrusion as well as existing one to intrusion detection system in the network environment.

• IMMUNE NETWORK
• /
• v.4 no.2
• /
• pp.81-87
• /
• 2004

Background: In the thymus, developing thymocytes continually interact with thymic epithelial cell components. Self MHC restriction of mature T cells are imposed in the thymus through interaction of immature double positive thymocytes and thymic cortical epithelial cells. The site of negative selection, however, is a matter of debate. Through systemic injection of anti-TCR antibody or antigenic peptides, investigators suggested that most of the negative selection occurs in the thymic cortex. But the requirements for negative selection, i.e cellular counterparts and costimulatory molecules are more available in the medulla or cortico-medullary junction rather than in the thymic cortex. Methods: The direct and indirect pathways of thymocyte death after systemic anti-TCR antibody injection were separated through several experimental systems. B6 mice were either adrenalectomized or sham-adrenalectomized to evaluate the role of endogenous glucocorticoids from adrenal gland. Role of TNF were evaluated through using TNF receptor double knockout mice. Results: We found that without indirectly acting mediators such as $TNF-\alpha$ or corticosteroid, double positive thymocyte death were minimal by systemic injection of anti-TCR antibody in TNF receptor double knockout neonatal mice. Also by analyzing neonatal wild-type mice with adoptively transferred mature T cells, only peripheral activation of mature T cells could induce extensive double positive thymocyte death. Conclusion: Thus, systemically injected anti-TCR antibody mediated thymocyte death are mostly induced through indirect pathway.

• Reproductive and Developmental Biology
• /
• v.38 no.2
• /
• pp.71-77
• /
• 2014

The specific genetic modification in porcine somatic cells by gene targeting has been very difficult because of low efficiency of homologous recombination. To improve gene targeting, we designed three kinds of knock-out vectors with ${\alpha}1,3$-galactosyltransferase gene (${\alpha}1,3$-GT gene), DT-A/pGT5'/neo/pGT3', DT-A/NLS/pGT5'/neo/pGT3' and pGT5'/neo/ pGT3'/NLS. The knock-out vectors consisted of a 4.8-kb fragment as the 5' recombination arm (pGT5') and a 1.9-kb fragment as the 3' recombination arm (pGT3'). We used the neomycin resistance gene (neo) as a positive selectable marker and the diphtheria toxin A (DT-A) gene as a negative selectable marker. These vectors have a neo gene insertion in exon 9 for inactivation of ${\alpha}1,3$-GT locus. DT-A/pGT5'/neo/pGT3' vector contain only positive-negative selection marker with conventional targeting vector. DT-A/NLS/pGT5'/neo/pGT3' vector contain positive-negative selection marker and NLS sequences in upstream of 5' recombination arm which enhances nuclear transport of foreign DNA into bovine somatic cells. pGT5'/neo/pGT3'/NLS vector contain only positive selection marker and NLS sequence in downstream of 3' recombination arm, not contain negative selectable marker. For transfection, linearzed vectors were introduced into porcine ear fibroblasts by electroporation. After 48 hours, the transfected cells were selected with $300{\mu}g/ml$ G418 during 12 day. The G418-resistant colonies were picked, of which 5 colonies were positive for ${\alpha}1,3$-GT gene disruption in 3' PCR and southern blot screening. Three knock-out somatic cells were obtained from DT-A/NLS/ pGT5'/neo/pGT3' knock-out vector. Thus, these data indicate that gene targeting vector using nuclear localization signal and negative selection marker improve targeting efficiency in porcine somatic cells.

• Smart Structures and Systems
• /
• v.20 no.2
• /
• pp.219-229
• /
• 2017

Outlier detection is an imperative task to identify the occurrence of abnormal events before the structures are suffered from sudden failure during their service lives. This paper proposes a two-phase method for the outlier detection of Global Positioning System (GPS) monitoring data. Prompt judgment of the occurrence of abnormal data is firstly carried out by use of the relational analysis as the relationship among the data obtained from the adjacent locations following a certain rule. Then, a negative selection algorithm (NSA) is adopted for further accurate localization of the abnormal data. To reduce the computation cost in the NSA, an improved scheme by integrating the adjustable radius into the training stage is designed and implemented. Numerical simulations and experimental verifications demonstrate that the proposed method is encouraging compared with the original method in the aspects of efficiency and reliability. This method is only based on the monitoring data without the requirement of the engineer expertise on the structural operational characteristics, which can be easily embedded in a software system for the continuous and reliable monitoring of civil infrastructure.

• IMMUNE NETWORK
• /
• v.15 no.3
• /
• pp.111-120
• /
• 2015

Dendritic cells (DCs) play a significant role in establishing self-tolerance through their ability to present self-antigens to developing T cells in the thymus. DCs are predominantly localized in the medullary region of thymus and present a broad range of self-antigens, which include tissue-restricted antigens expressed and transferred from medullary thymic epithelial cells, circulating antigens directly captured by thymic DCs through coticomedullary junction blood vessels, and peripheral tissue antigens captured and transported by peripheral tissue DCs homing to the thymus. When antigen-presenting DCs make a high affinity interaction with antigen-specific thymocytes, this interaction drives the interacting thymocytes to death, a process often referred to as negative selection, which fundamentally blocks the self-reactive thymocytes from differentiating into mature T cells. Alternatively, the interacting thymocytes differentiate into the regulatory T (Treg) cells, a distinct T cell subset with potent immune suppressive activities. The specific mechanisms by which thymic DCs differentiate Treg cells have been proposed by several laboratories. Here, we review the literatures that elucidate the contribution of thymic DCs to negative selection and Treg cell differentiation, and discusses its potential mechanisms and future directions.

• Journal of Information Processing Systems
• /
• v.2 no.1
• /
• pp.23-27
• /
• 2006

This paper presents a feature selection technique based on distributional differences for efficient machine learning. Initial training data consists of data including many features and a target value. We classified them into positive and negative data based on the target value. We then divided the range of the feature values into 10 intervals and calculated the distribution of the intervals in each positive and negative data. Then, we selected the features and the intervals of the features for which the distributional differences are over a certain threshold. Using the selected intervals and features, we could obtain the reduced training data. In the experiments, we will show that the reduced training data can reduce the training time of the neural network by about 40%, and we can obtain more profit on simulated stock trading using the trained functions as well.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2002.12a
• /
• pp.478-481
• /
• 2002

본 논문에서는 생명체의 면역계에서 중요한 역할을 하는 세포독성 T세포의 생성과정의 하나인 긍정선택(positive selection)과 부정 선택(negative selection)을 모델링하여 침입에 의한 데이터 변경과 바이러스에 의한 데이터 감염 등을 탐지할 때 가장 중요한 요소인 변경 검사 알고리즘을 개발하였다. 제안한 알고리즘은 면역세포의 생성시 MHC 인식부를 형성해 주는 긍정 선택을 자기 인식 알고리즘으로 구현하여 컴퓨터에서 자기로 인식해야하는 파일이나 기능에 대해 MHC 인식부를 형성하고, 또한 항원 인식부를 형성하는 부정 선택을 이용해 변형 검지기(anomaly detector)를 구성한다. 따라서 제안한 알고리즘은 실제 면역세포와 마찬가지로 자신과 침입자 모두에 대한 인식기를 가지고 변경을 탐지하게 된다. 시뮬레이션을 통하여 자기파일의 일부가 변경되었을 때와 블록이 변경되었을 때에 대하여 두 가지 방법을 이용한 변경 검사 알고리즘의 특성과 유효성을 밝힌다.