• 디지털산업정보학회논문지
• /
• 제13권4호
• /
• pp.125-139
• /
• 2017

In this paper, we developed a framework to detect and predict insider information leakage by collecting and restoring network traffic. For automated behavior analysis, many meta information and behavior information obtained using network traffic collection are used as machine learning features. By these features, we created and learned behavior model, network model and protocol-specific models. In addition, the ensemble model was developed by digitizing and summing the results of various models. We developed a function to present information leakage candidates and view meta information and behavior information from various perspectives using the visual analysis. This supports to rule-based threat detection and machine learning based threat detection. In the future, we plan to make an ensemble model that applies a regression model to the results of the models, and plan to develop a model with deep learning technology.

• ETRI Journal
• /
• 제46권1호
• /
• pp.59-70
• /
• 2024

We introduce a high-performance named entity recognition (NER) model for written and spoken language. To overcome challenges related to labeled data scarcity and domain shifts, we use transfer learning to leverage our previously developed KorBERT as the base model. We also adopt a meta-pseudo-label method using a teacher/student framework with labeled and unlabeled data. Our model presents two modifications. First, the student model is updated with an average loss from both human- and pseudo-labeled data. Second, the influence of noisy pseudo-labeled data is mitigated by considering feedback scores and updating the teacher model only when below a threshold (0.0005). We achieve the target NER performance in the spoken language domain and improve that in the written language domain by proposing a straightforward rollback method that reverts to the best model based on scarce human-labeled data. Further improvement is achieved by adjusting the label vector weights in the named entity dictionary.

• Asian Pacific Journal of Cancer Prevention
• /
• 제15권13호
• /
• pp.5277-5281
• /
• 2014

Published studies have evaluated associations between the MDM2 SNP309T>G polymorphism and bladder cancer susceptibility. However, these generated inconsistent results. The aim of the present investigation was to quantify the strength of association between MDM2 SNP309T>G polymorphism and bladder cancer risk by conducting a meta-analysis. We searched PubMed and Embase for related studies that had been published in English before April 1, 2014 and associations were assessed by summarizing the odds ratios (ORs) with the corresponding 95% confidence intervals (CIs). Five case-control studies with a total of 972 cases and 1,012 controls were finally identified to be eligible for the meta-analysis. Overall, the results indicated that there was no significant association between the MDM2 SNP309T>G polymorphism and bladder cancer risk (for the allele model G vs. T: OR=1.08, 95% CI 0.85-1.36, p=0.54; for the co-dominant model GG vs. TT: OR=1.20, 95% CI 0.74-1.93, p=0.46; for the dominant model GG+GT vs. TT: OR=0.98, 95% CI 0.80-1.20, p=0.83; for the recessive model GG vs. GT+TT: OR=1.20, 95% CI 0.83-1.74, p=0.33). However, on subgroup analysis by ethnicity, significant associations were found in Caucasians in three models (for the allele model G vs. T: OR=1.41, 95% CI 1.10-1.81, p=0.006; for the co-dominant model GG vs. TT: OR=2.16, 95% CI 1.28-3.63, p=0.004; for the recessive model GG vs. GT+TT: OR=2.06, 95% CI 1.31-3.22, p=0.002). In summary, the present meta-analysis provides evidence that the genotype for the MDM2 SNP309T>G polymorphism may be associated with genetic susceptibility to bladder cancer among Caucasians.

• 디지털융복합연구
• /
• 제11권5호
• /
• pp.437-442
• /
• 2013

최근에는 IT와 관광을 융합하는 IT관광 융합학문이 미래학문으로 떠오르고 있다. 본 연구는 u-스마트 관광정보 시스템을 이용한 콘텐츠 정보품질에 대한 사항들을 설계 및 구현한다. 이것은 관광정보 패턴 템플리트 라이브러리에 있는 콘텐츠 정보품질 사항을 구현함으로써 관광객의 관광성향과 관광평점을 분석할 수 있는 기초 프로세스와 데이터를 생성할 수 있다. 본 논문은 스마트 앱과 XMI를 활용하여 관광정보 콘텐츠 정보품질에 맞는 콘텐츠를 생성하고 구현함으로써 관광정보와 관광객의 성향을 실시간으로 알 수 있는 것이 장점이다. 뿐만 아니라, 스마트 앱으로 메타모델 연결관계 상호 관련기 프로세스를 거쳐 메타 모델화할 수 있다. 관광정보 콘텐츠의 정보품질을 구현함으로써 관광객의 성향을 파악할 수 있으며 향후 관광디지털정책을 디자인할 수 있는 모형개발에 도움을 준다.

• Asian Pacific Journal of Cancer Prevention
• /
• 제15권22호
• /
• pp.9643-9647
• /
• 2014

Background: Several studies have been performed to investigate the association of the HER2 Ile655Val polymorphism and breast cancer risk. However, the results were inconsistent. To understand the precise relationship, a meta-analysis was here conducted. Materials and Methods: A search of PubMed conducted to investigate links between the HER2 Ile655Val polymorphism and breast cancer, identified a total of 32 studies, of which 29, including 14,926 cases and 15,768 controls, with odds ratios (ORs) with 95% confidence intervals were used to assess any association. Results: In the overall analysis, the HER2 Ile655Val polymorphism was associated with breast cancer in an additive genetic model (OR=1.136, 95% CI 1.043-1.239, p=0.004) and in a dominant genetic (OR=1.118, 95% CI 1.020-1.227, p=0.018), while no association was found in a recessive genetic model. On subgroup analysis, an association with breast cancer was noted in the additive genetic model (OR=1.111, 95% CI: 1.004-1.230, p=0.042) for the Caucasian subgroup. No significant associations were observed in Asians and Africans in any of the genetic models. Conclusions: In summary, our meta-analysis findings suggest that the HER2 Ile655Val polymorphism is marginally associated with breast cancer susceptibility in worldwide populations with additive and dominant models, but not a recessive model.

• 한국멀티미디어학회논문지
• /
• 제12권2호
• /
• pp.231-243
• /
• 2009

임베디드 시스템 시장이 확대되면서 최근 들어 자동차, 선박, 로봇 등과 같은 다양한 분야에서 임베디드 소프트웨어에 대한 관심이 증가되고 있다. 이와 관련하여 임베디드 운영체제, 임베디드 소프트웨어 모델링 기법, 임베디드 소프트웨어 테스팅 등 다양한 연구들이 이루어지고 있다. 그러나 지금까지 임베디드 분야가 하드웨어 분야에 치중되어 왔기 때문에 임베디드 시스템 개발에 있어서 체계적인 공학적 접근 방식이 미약한 상태이다. 이 뿐만 아니라 임베디드 시스템 개발에 있어서 재사용성을 고려한 프레임워크 기반의 설계 기법 등은 반영되지 못한 실정이다. 이렇게 개발됨으로써 시스템 내에 죽은 코드(Dead Code)들이 산재해 있을 뿐만 아니라 시스템의 재사용성이 매우 저조한 결과를 보이고 있다. 본 논문에서는 이러한 재사용성 향상을 위한 임베디드 시스템의 프레임워크를 제시하고, 이 프레임워크에 대한 정적 메타 모델을 제시한다. 이러한 메타 모델을 통해 임베디드 시스템의 재사용 프레임워크의 구조를 표현 할 뿐만 아니라, 이를 기반으로 다양한 임베디드 시스템 형태에 따라 모델을 쉽게 확장하여 설계할 수 있도록 한다.

• 한국빅데이터학회지
• /
• 제2권2호
• /
• pp.47-56
• /
• 2017

통합기술수용모형(UTAUT)은 기술수용모형(TAM, Technology Acceptance Model)의 한계점을 개선하기 위해 8가지의 기존 이론을 종합하여 제시된 것으로, 최근 신기술의 수용 및 확산과 관련된 다양한 분야에서 활용되어지고 있다. 본 연구는 메타분석을 통하여 UTAUT를 이용한 연구의 주요요인에 대해 분석과 연구모형의 일관성에 대한 검증을 하였다. 이를 통해 주요요인의 관계와 추가적 요인에 대한 결과를 제시하고 해석하였다. 아울러 선행연구와의 비교분석을 통해 차이점과 시사점을 제시하였다. 메타분석 결과, 추가적 요인으로 만족감, 쾌락동기, 태도, 지각된 즐거움이 중요한 것으로 분석되었으며, 이 결과를 바탕으로 확장된 UTAUT 모형을 제시하였다. 또한, 국내 연구의 경우 행동의도를 높이는 활동이 실제 행동으로 이어지는데 가장 중요한 요인으로 나타났다. 본 연구의 결과는 신기술이나 신제품의 수용 및 확산을 연구하는 연구자 및 신제품을 출시하고자 하는 기업에 도움을 제공할 것으로 예상한다.

• Asian Pacific Journal of Cancer Prevention
• /
• 제13권8호
• /
• pp.3763-3766
• /
• 2012

Background: Methylenetetrahydrofolate reductase (MTHFR) is a key enzyme in the metabolism of folate, and the role of the MTHFR C677T polymorphism in pancreatic carcinogenesis is still controversial. Methods: A literature search was performed using Pubmed and CNKI databases for published studies through May 2012. We performed a meta-analysis of all relevant case-control studies that examined the association between MTHFR C677T polymorphism and pancreatic cancer risk. Results: Finally, 9 individual case-control studies with a total of 1,299 pancreatic cancer cases and 2,473 controls were included into this meta-analysis. Results: This metaanalysis showed there was an obvious association between MTHFR C677T polymorphism and pancreatic cancer risk in East Asians (for allele model, OR = 1.67, 95%CI 1.11-2.51; For homozygote model, OR = 2.77, 95%CI 1.40-5.48; for recessive model, OR = 1.96, 95%CI 1.54-2.50; for dominant model, OR = 2.11, 95%CI 1.01-4.41). However, no significant association was found in Caucasians. Conclusions: The MTHFR C677T polymorphism is associated with pancreatic cancer risk, and a race-specific effect may exist in this association. More studies with a larger sample size are needed to further clarify this association.

• Asian Pacific Journal of Cancer Prevention
• /
• 제15권17호
• /
• pp.7251-7255
• /
• 2014

MicroRNAs (miRNAs) act as tumor suppressors or promoters in neoplasia by regulating relative geneexpression. The association between a single nucleotide polymorphism (SNP) rs4938723 in miR-34b/c and susceptibility to cancers was inconsistent in previous studies. In this study, we conducted a literature search of PubMed, Web of Science and Embase to identify all relevant studies in this meta-analysis with 6,036 cases and 6,204 controls. We found that the miR-34b/c rs4938723 polymorphism was significantly associated with increased risk of cancers in the heterozygous model (TC versus TT, OR=1.09, 95% CI=1.01-1.18, P=0.02). Subgroup analysis also revealed increased risk for Asian ethnicity in the heterozygous model (TC versus TT, OR=1.12, 95% CI=1.02-1.22, P=0.02), but decreased risk of colorectal cancer in homozygote model (CC versus TT, OR=0.66, 95% CI=0.47-0.92, P=0.02) and in the recessive model (CC versus TC+TT, OR=0.67, 95% CI=0.48-0.93, P=0.02) by cancer type. The current meta-analysis indicated that the miR-34b/c rs4938723 polymorphism may decrease susceptibility to colorectal cancer. Well-designed studies with larger sample size are required to further validate the results.

• Asian Pacific Journal of Cancer Prevention
• /
• 제13권5호
• /
• pp.2325-2328
• /
• 2012

Background: Cyclin D1 (CCND1) is critical in the transition of the cell cycle from G1 to S phases and unbalanced cell cycle regulation is a hallmark of carcinogenesis. A number of studies conducted to assess the association between CCND1 G870A polymorphism and susceptibility to lung cancer have yielded inconsistent and inconclusive results. In the present study, the possible association above was assessed by a meta-analysis. Methods: Eligible articles were identified for the period up to November 2011. Pooled odds ratios (OR) with 95% confidence intervals (95%CI) were appropriately derived from fixed effects or random-effects models. Sensitivity analysis excluding studies whose genotype frequencies in controls significantly deviated from the Hardy-Weinberg equilibrium (HWE) was performed. Results: Ten case-control studies with a total of 10,548 subjects were eligible. At the overall analysis the CCND1 870A allele appeared to be associated with elevated lung cancer risk (for allele model, pooled OR = 1.24, 95% CI: 1.08-1.44, P = 0.004; for homozygous model, pooled OR = 1.45, 95% CI: 1.14-1.84, P = 0.003; for recessive model, pooled OR = 1.29, 95% CI: 1.06-1.58, P = 0.013; for dominant model, pooled OR = 1.33, 95% CI: 1.08-1.65, P = 0.009). Subgroup analyses by ethnicity and sensitivity analysis further pointed to associations, particularly in Asians. Conclusion: This meta-analysis suggests that the A allele of CCND1 G870A polymorphism confers additional lung cancer risk.