• The KIPS Transactions:PartC
• /
• v.12C no.4 s.100
• /
• pp.609-616
• /
• 2005

Signature-based Intrusion Detection has many false positive and many difficulties to detect new and changed attacks. Alpha-cut is introduced which reduces false positive with a combination of signature-based IDS and machine learning-based IDS in prior paper [1]. This research is a study of a succession of Alpha-cut, and we introduce Beta-rick in which attacks can be detected but cannot be detected in single signature-based detection. Alpha-cut is a way of increasing detection accuracy for the signature based IDS, Beta-pick is a way which decreases the case of treating attack as normality. For Alpha-cut and Beta-pick we use XIBL as a learning algorithm and also show the difference of result of Sd.5. To describe the value of proposed method we apply Alpha-cut and Beta-pick to signature-based IDS and show the decrease of false alarms.

• 한국어정보학회:학술대회논문집
• /
• 2016.10a
• /
• pp.333-336
• /
• 2016

본 연구에서는 개체명 인식의 성능을 향상시키기 위해, 가중 투표 방법을 이용하여 개체명 인식 모델을 앙상블 하는 방법을 제안한다. 각 모델은 Conditional Random Fields의 변형 알고리즘을 사용하여 학습하고, 모델들의 가중치는 다목적 함수 최적화 기법인 NSGA-II 알고리즘으로 학습한다. 실험 결과 제안 시스템은 $F_1Score$ 기준으로 87.62%의 성능을 보여, 단독 모델 중 가장 높은 성능을 보인 방법보다 2.15%p 성능이 향상되었다.

• Annual Conference on Human and Language Technology
• /
• 2016.10a
• /
• pp.333-336
• /
• 2016

본 연구에서는 개체명 인식의 성능을 향상시키기 위해, 가중 투표 방법을 이용하여 개체명 인식 모델을 앙상블 하는 방법을 제안한다. 각 모델은 Conditional Random Fields의 변형 알고리즘을 사용하여 학습하고, 모델들의 가중치는 다목적 함수 최적화 기법인 NSGA-II 알고리즘으로 학습한다. 실험 결과 제안 시스템은 $F_1Score$기준으로 87.62%의 성능을 보여, 단독 모델 중 가장 높은 성능을 보인 방법보다 2.15%p 성능이 향상되었다.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2022.07a
• /
• pp.535-536
• /
• 2022

본 논문 원격교육환경에서 학습자가 남긴 개별 데이터에 대한 장기적 관리 및 효율적 학습자 관리를 위한 데이터 전처리 방법으로 전통적인 ARIMA를 수정하여 연구하였다. ARIMA는 과거시점 데이터에 대한 회귀식과 변화율을 현 시점 데이터에 반영하는 방식이며 본 연구에서는 ARIMA 처리과정에서 딥러닝 알고리즘인 RNN의 변형방법인 LSTM을 적용하여 부분 데이터셋의 전처리과정에 대한 정확성과 재현율을 높이도록 하였다. 본 연구의 결과 전통적인 ARIMA 적용시와 대비하여 7~9%의 성능향상을 확인하였다.

• Journal of the Korea institute for structural maintenance and inspection
• /
• v.24 no.6
• /
• pp.206-216
• /
• 2020

As the number of aging railway bridges in Korea increases, maintenance costs due to aging are increasing and continuous management is becoming more important. However, while the number of old facilities to be managed increases, there is a shortage of professional personnel capable of inspecting and diagnosing these old facilities. To solve these problems, this study presents an improved model that can detect Local damage to structures using machine learning techniques of AI technology. To construct a damage detection machine learning model, an analysis model of the bridge was set by referring to the design drawing of a non-ballasted plate-girder railroad bridge. Static strain data according to the damage scenario was extracted with the analysis model, and the Local damage index based on the reliability of the bridge was presented using statistical techniques. Damage was performed in a three-step process of identifying the damage existence, the damage location, and the damage severity. In the estimation of the damage severity, a linear regression model was additionally considered to detect random damage. Finally, the random damage location was estimated and verified using a machine learning-based damage detection classification learning model and a regression model.

• Annual Conference on Human and Language Technology
• /
• 2014.10a
• /
• pp.95-99
• /
• 2014

본 논문에서는 최근 급속히 증가하여 사회적 이슈가 되고 있는 SMS 스팸 필터링을 위한 듀얼 SMS 스팸필터링 기법을 제안한다. 지속적으로 증가하고 새롭게 변형되는 SMS 문자 필터링을 위해서는 패턴 및 스팸 단어 사전을 통한 필터링은 많은 수작업을 요구하여 부적합하다. 그리하여 기계 학습을 이용한 자동화 시스템 구축이 요구되고 있으며, 효과적인 기계 학습을 위해서는 자질 선택과 자질의 가중치 책정 방법이 중요하다. 하지만 SMS 문자 특성상 문장들이 짧기 때문에 출현하는 자질의 수가 적어 분류의 어려움을 겪게 된다. 이 같은 문제를 개선하기 위하여 본 논문에서는 슬라이딩 윈도우 기반 N-gram 확장을 통해 자질을 확장하고, 확장된 자질로 그래프를 구축하여 얕은 구조적 특징을 표현한다. 학습 데이터에 출현한 N-gram 자질을 정점(Vertex)으로, 자질의 출현 빈도를 그래프의 간선(Edge)의 가중치로 설정하여 햄(HAM)과 스팸(SPAM) 그래프를 각각 구성한다. 이렇게 구성된 그래프를 바탕으로 노드의 중요도와 간선의 가중치를 활용하여 최종적인 자질의 가중치를 결정한다. 입력 문자가 도착하면 스팸과 햄의 그래프를 각각 이용하여 입력 문자의 2개의 자질 벡터(Vector)를 생성한다. 생성된 자질 벡터를 지지 벡터 기계(Support Vector Machine)를 이용하여 각 SVM 확률 값(Probability Score)을 얻어 스팸 여부를 결정한다. 3가지의 실험환경에서 바이그램 자질과 이진 가중치를 사용한 기본 시스템보다 F1-Score의 약 최대 2.7%, 최소 0.5%까지 향상되었으며, 결과적으로 평균 약 1.35%의 성능 향상을 얻을 수 있었다.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.36S no.10
• /
• pp.103-112
• /
• 1999

Genetic algorithm, which uses principles of natural selection and population genetics, is an efficient method to find out an optimal solution. In conventional genetic algorithms, however, the size of gene pool needs to be increased to insure a convergency. Therefore, many memory spaces and much computation time were needed. Also, since child chromosomes were generated by chromosome crossover and gene mutation, the algorithms have a complex structure. Thus, in this paper, a compact stereo matching algorithm using a population-based incremental learning based on probability vector is proposed to reduce these problems. The PBIL method is modified for matching environment. Since th proposed algorithm uses a probability vector and eliminates gene pool, chromosome crossover, and gene mutation, the matching algorithm is simple and the computation load is considerably reduced. Even though the characteristics of images are changed, stable outputs are obtained without the modification of the matching algorithm.

• The Journal of Korean Association of Computer Education
• /
• v.12 no.5
• /
• pp.35-47
• /
• 2009

Just as noted in some studies related to U-learning based on the ubiquitous technology, it is required to reflect both factors of the environments and contexts of learners on the teaching and learning strategy. For this, the current paper suggests an inquiry learning method and a mobile learning support device for the inquiry learning. Since the inquiry learning is operating on the basis of the contents that learners themselves experience in the fields, it naturally leads them to position in the learning environments and contexts. In this study some curricula are adopted and transformed for the inquiry learning, and the developed mobile device for the learning has a guide function, so that learners might follow the inquiry learning process. In addition, it is possible to reuse the acquired learning data by storing them on a remote learning sever, which may support the inquiry learning of students. These system and device are applied in the teaching sites of schools and, consequently, it is found that in regard with the learning interest and academic performance of students there is an advanced result, being statistically significant.

• Journal of Internet Computing and Services
• /
• v.22 no.2
• /
• pp.1-9
• /
• 2021

Traditionally, most malicious codes have been analyzed using feature information extracted by domain experts. However, this feature-based analysis method depends on the analyst's capabilities and has limitations in detecting variant malicious codes that have modified existing malicious codes. In this study, we propose a ResNet-Variational AutoEncder-based variant malware classification method that can classify a family of variant malware without domain expert intervention. The Variational AutoEncoder network has the characteristics of creating new data within a normal distribution and understanding the characteristics of the data well in the learning process of training data provided as input values. In this study, important features of malicious code could be extracted by extracting latent variables in the learning process of Variational AutoEncoder. In addition, transfer learning was performed to better learn the characteristics of the training data and increase the efficiency of learning. The learning parameters of the ResNet-152 model pre-trained with the ImageNet Dataset were transferred to the learning parameters of the Encoder Network. The ResNet-Variational AutoEncoder that performed transfer learning showed higher performance than the existing Variational AutoEncoder and provided learning efficiency. Meanwhile, an ensemble model, Stacking Classifier, was used as a method for classifying variant malicious codes. As a result of learning the Stacking Classifier based on the characteristic data of the variant malware extracted by the Encoder Network of the ResNet-VAE model, an accuracy of 98.66% and an F1-Score of 98.68 were obtained.

• Journal of Internet Computing and Services
• /
• v.20 no.6
• /
• pp.137-142
• /
• 2019

This paper Various services exist to detect and monitor abnormal event. However, most services focus on fires and gas leaks. so It is impossible to prevent and respond to emergency situations for the elderly and severely disabled people living alone. In this study, AI model is designed and compared to detect abnormal event of heart rate signal which is considered to be the most important among various bio signals. Specifically, electrocardiogram (ECG) data is collected using Physionet's MIT-BIH Arrhythmia Database, an open medical data. The collected data is transformed in different ways. We then compare the trained AI model with the modified and ECG data.