• 대한통합의학회지
• /
• 제10권4호
• /
• pp.145-151
• /
• 2022

Purpose : Stroke patients exhibit abnormal walking patterns such as slow walking speed and asymmetrical walking values. The recovery of symmetrical walking in the stance phase using a treadmill means improvements in walking speed and asymmetrical walking. The purpose of this research was to investigate the effect of unilateral step treadmill training (USTT) on gait speed and the recovery of symmetrical walking in chronic stroke patients. Methods : Fifteen patients (11 men and 4 women) with chronic stroke participated in this study. The 10-meter walk test (10MWT) and GAITRite system were used to determine the intervention-related changes in gait speed and symmetrical walking values such as non-paretic step length (NSL), non-paretic step time (NST), paretic single-support time (PSST), step length asymmetry (SLA), and step time asymmetry (STA) after USTT. All participants completed USTT and underwent measurements at 3 different times: at pretest, posttest, and the follow-up test. Repeated-measures analysis of variance was used to compare walking speed and asymmetrical walking values. The statistical significance level was set at p<.05. Results : Walking speed by 10MWT (p<.05) showed significant improvements after USTT as follows: at pretest and posttest (p<.05), posttest and follow-up test (p<.05), and pretest and follow-up test (p<.05). Recovery of symmetrical walking patterns such as NSL (p<.05), NST (p<.05), and SLA (p<.05) were observed after USTT. However, no significant improvements were found in PSST (p>.05) and STA (p>.05) in symmetrical gait. Conclusion : This study suggests that USTT may have a positive effect on walking speed and symmetrical walking patterns in chronic stroke patients. Thus, this study contributes to the existing knowledge about the usefulness of USTT for the effective management of patients with chronic stroke. Further studies are needed to generalize these findings.

• International Journal of Computer Science & Network Security
• /
• 제23권9호
• /
• pp.55-64
• /
• 2023

The Internet of Things (IoT) is the combination of the internet and various sensing devices. IoT security has increasingly attracted extensive attention. However, significant losses appears due to malicious attacks. Therefore, intrusion detection, which detects malicious attacks and their behaviors in IoT devices plays a crucial role in IoT security. The intrusion detection system, namely IDS should be executed efficiently by conducting classification and efficient feature extraction techniques. To effectively perform Intrusion detection in IoT applications, a novel method based on a Conventional Neural Network (CNN) for classification and an improved Genetic Algorithm (GA) for extraction is proposed and implemented. Existing issues like failing to detect the few attacks from smaller samples are focused, and hence the proposed novel CNN is applied to detect almost all attacks from small to large samples. For that purpose, the feature selection is essential. Thus, the genetic algorithm is improved to identify the best fitness values to perform accurate feature selection. To evaluate the performance, the NSL-KDDCUP dataset is used, and two datasets such as KDDTEST²¹ and KDDTEST⁺ are chosen. The performance and results are compared and analyzed with other existing models. The experimental results show that the proposed algorithm has superior intrusion detection rates to existing models, where the accuracy and true positive rate improve and the false positive rate decrease. In addition, the proposed algorithm indicates better performance on KDDTEST⁺ than KDDTEST²¹ because there are few attacks from minor samples in KDDTEST⁺. Therefore, the results demonstrate that the novel proposed CNN with the improved GA can identify almost every intrusion.

• 한국정보통신학회논문지
• /
• 제26권12호
• /
• pp.1872-1879
• /
• 2022

최근 인공지능 기술이 발전하면서 해킹 공격을 탐지하기 위해 인공지능을 이용하려는 연구가 활발히 진행되고 있다. 하지만, 인공지능 모델 개발에 핵심인 학습데이터를 구성하는데 있어서 보안데이터가 대표적인 불균형 데이터라는 점이 큰 장애물로 인식되고 있다. 이에 본 눈문에서는 오버샘플링을 위한 데이터 추출에 딥러닝 생성 모델인 VAE를 적용하고 K-NN을 이용한 가중치 계산을 통해 클래스별 오버샘플링 개수를 설정하여 샘플링을 하는 W-VAE 오버샘플링 기법을 제안한다. 본 논문에서는 공개 네트워크 보안 데이터셋인 NSL-KDD를 통해 ROS, SMOTE, ADASYN 등 총 5가지 오버샘플링 기법을 적용하였으며 본 논문에서 제안한 오버샘플링 기법이 F1-Score 평가지표를 통해 기존 오버샘플링 기법과 비교하여 가장 효과적인 샘플링 기법임을 증명하였다.

• 대한치과교정학회지
• /
• 제27권5호
• /
• pp.743-760
• /
• 1997

두개저의 크기, 형태 및 두부자세가 지니는 두개안면구조의 공간적 수평, 수직적 위치간의 상관성을 파악해 보고자 남자 51명, 여자 49명으로 구성된 표본으로부터 촬영된 100장의 측모 두부방사선사진을 이용, 12개의 계측항목과 37개의 기준점을 설정하고 계측항목에 대한 계측치와 기준점의 수평, 수직 위치를 산출한 다음,이들 간의 상관관계를 통계적으로 분석하였으며 계측항목에서 얻은 계측치의 크기가 크거나 작은 군을 각각 10개 표본씩으로 분류하여 mean facial diagram을 작성, 비교함으로서 다음과 같은 결론을 얻었다. 1. 두개저의 형태변수인 n-s-ba 및 n-s-ar각은 경추의 기준점 cv4ip, cv2ip, cv4tg 그리고 cv2ap의 수평, 수직적 위치 모두에 높은 통계적 유의성의 상관관계를 보였다($0.1\%$ 유의수준). 2. 두개저의 형태변수인 n-s-ba및 n-s-ar각은 안면구조에 있는 대부분의 기준점의 수평적 위치와 상관관계를 보였으나($1\%$ 유의수준), 수직적 위치는 통계적 유의성이 없었다($5\%$ 유의수준). 3. 두개저의 크기변수인 n-s, n-ba, n-ba및 n-ar의 크기는 두개안면구조내 기준점의 위치와 다양한 양상의 상관관계를 보였으나, 대체로 치아, 치조와 관련된 중안면구조의 수평, 수직적 위치와 상관관계가 높았다. 4. 두개저와 경추의 기울기가 이루는 자세변수인 NSL/CVT, NSL/OPT각은 두개안면구조내 기준점의 수평적 위치에 높은 상관성을 지니고 있었으며 통계학적 유의성이 인정되었으나($1\%$유의수준), 수직적 위치와는 통계적 유의성이 없었다($5\%$ 유의수준). 5. 진수평선(true horizontal line)과 경추의 기울기가 이루는 자세변수인 OPT/HOR 및 CVT/HOR각은 두개안면구조내 기준점의 수평, 수직적 위치 모두와 통계적으로 유의성이 인정되는 상관성을 보이지 않았다($5\%$ 유의수준). 6. 연조직에 존재하는 기준점의 수평, 수직적 위치는 모든 변수와의 상관성에서 대부분 경조직의 관련 기준점에 준하는 양상을 보였다.

• Corrosion Science and Technology
• /
• 제6권4호
• /
• pp.211-218
• /
• 2007

In this paper, evaluation and development of corrosion resistant materials for smokestacks is summarized mainly on the basis of the author's experimental results. Operating environments of smokestacks and the problems of conventional lining materials for smokestacks are described briefly. The emphasis is focused upon the evaluation and development of recently developed corrosion resistant steels such as YUS260 for heavy oil fired smokestacks, WELACC5 for LNG fired smokestacks and NSL310MoCu Clad steel for coal fired smokestacks. Corrosion resistance of these steels under laboratory corrosion testing environments and actual environments are evaluated. Finally future problems of corrosion resistant materials for smokestacks are touched on briefly.

• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2002년도 학술발표회 논문집(I)
• /
• pp.27-40
• /
• 2002

Developed at Colorado State University, CASC2D-SED is a physically-based model simulating the hydrologic response of a watershed to a distributed rainfall field. The time-dependent processes include: precipitation, interception, infiltration, surface runoff and channel routing, upland erosion, transport and sedimentation. CASC2D-SED is applied to Goodwin Creek, Mississippi. The watershed covers 21 $\textrm{km}^2$ and has been extensively monitored both at the outlet and at several internal locations by the ARS-NSL at Oxford, MS. The model has been calibrated and validated using rainfall data from 16 meteorological stations,6 stream gaging stations and 6 sediment gaging stations. Sediment erosion/deposition rates by size fraction are predicted both in space and time. Geovisualization, a powerful data exploration technique based on GIS technology, is used to analyze and display the dynamic output time series generated by the CASC2D-SED model.

• 한국전기전자재료학회논문지
• /
• 제26권3호
• /
• pp.227-231
• /
• 2013

Electrodeionization is a hybrid separation process of electrodialysis and ion exchange to produce high purity water under electric field. This article provides a fabrication result of hole patterned metal electrode for elecrodeionization system. The hole patterns have been fabricated by nanosphere lithography (NSL). The technique utilizes the self-assembled nanospheres as lens-mask patterns and collimated laser beam source. The hole patterns have a periodic array structure. The images of hole pattern on metal electrode prepared were observed by SEM. We believe that the periodic hole patterned metal electrode structure is a useful device applicable for metal mat electrode in electrodeionization system.

• ETRI Journal
• /
• 제43권1호
• /
• pp.152-162
• /
• 2021

In the last few years, detection has become a powerful methodology for network protection and security. This paper presents a new detection scheme for data recorded over a computer network. This approach is applicable to the broad scientific field of information security, including intrusion detection and prevention. The proposed method employs bidimensional (time-frequency) data representations of the forms of the short-time Fourier transform, as well as the Wigner distribution. Moreover, the method applies matrix factorization using singular value decomposition and principal component analysis of the two-dimensional data representation matrices to detect intrusions. The current scheme was evaluated using numerous tests on network activities, which were recorded and presented in the KDD-NSL and UNSW-NB15 datasets. The efficiency and robustness of the technique have been experimentally proved.

• International Journal of Computer Science & Network Security
• /
• 제22권5호
• /
• pp.97-102
• /
• 2022

Network intrusion detection is becoming an increasing necessity for both organizations and individuals alike. Detecting intrusions is one of the major components that aims to prevent information compromise. Automated systems have been put to use due to the voluminous nature of the domain. The major challenge for automated models is the noise and data imbalance components contained in the network transactions. This work proposes an ensemble model, Attribute Subset Selector Bagging (ASUB) that can be used to effectively handle noise and data imbalance. The proposed model performs attribute subset based bag creation, leading to reduction of the influence of the noise factor. The constructed bagging model is heterogeneous in nature, hence leading to effective imbalance handling. Experiments were conducted on the standard intrusion detection datasets KDD CUP 99, Koyoto 2006 and NSL KDD. Results show effective performances, showing the high performance of the model.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2020년도 춘계학술발표대회
• /
• pp.526-529
• /
• 2020

최근 네트워크 환경에 대한 공격이 급속도로 고도화 및 지능화 되고 있기에, 기존의 시그니처 기반 침입탐지 시스템은 한계점이 명확해지고 있다. 이러한 문제를 해결하기 위해서 기계학습 기반의 침입 탐지 시스템에 대한 연구가 활발히 진행되고 있지만 기계학습을 침입 탐지에 이용하기 위해서는 두 가지 문제에 직면한다. 첫 번째는 실시간 탐지를 위한 학습과 연관된 중요 특징들을 선별하는 문제이며 두 번째는 학습에 사용되는 데이터의 불균형 문제로, 기계학습 알고리즘들은 데이터에 의존적이기에 이러한 문제는 치명적이다. 본 논문에서는 위 제시된 문제들을 해결하기 위해서 Hybrid Feature Selection과 Data Balancing을 통한 심층 신경망 기반의 네트워크 침입 탐지 모델을 제안한다. NSL-KDD 데이터 셋을 통해 학습을 진행하였으며, 평가를 위해 Accuracy, Precision, Recall, F1 Score 지표를 사용하였다. 본 논문에서 제안된 모델은 Random Forest 및 기본 심층 신경망 모델과 비교해 F1 Score를 기준으로 7~9%의 성능 향상을 이루었다.