• 인터넷정보학회논문지
• /
• 제22권1호
• /
• pp.13-22
• /
• 2021

최근 네트워크 환경에 대한 공격이 급속도로 고도화 및 지능화 되고 있기에, 기존의 시그니처 기반 침입탐지 시스템은 한계점이 명확해지고 있다. 지능형 지속 위협(Adavanced Persistent Threat; APT)과 같은 새로운 공격에 대해서 시그니처 패턴은 일반화 성능이 떨어지는 문제가 존재한다. 이러한 문제를 해결하기 위해 기계학습 기반의 침입 탐지 시스템에 대한 연구가 활발히 진행되고 있다. 하지만 실제 네트워크 환경에서 공격 샘플은 정상 샘플에 비해서 매우 적게 수집되어 클래스 불균형(Class Imbalance) 문제를 겪게 된다. 이러한 데이터로 지도 학습 기반의 이상 탐지 모델을 학습시킬 경우 정상 샘플에 편향된 결과를 가지게 된다. 본 논문에서는 이러한 불균형 문제를 해결하기 위해서 오토 인코더(Auto Encoder; AE)를 활용해 One-Class Anomaly Detection 을 수행하여 이를 극복한다. 실험은 NSL-KDD 데이터 셋을 통해 진행되었으며, 제안한 방법의 성능 평가를 위해 지도 학습된 모델들과 성능을 비교한다.

• 대한치과교정학회지
• /
• 제2권1호
• /
• pp.35-40
• /
• 1971

Since malocclusion affects a large segment of the population, it is by definition a public health problem. The etiology ana treatment of malocclusions have been studied by clinicians; however epidemioloic aspect of tile problem have been neglected. This study was undertaken using Angle's classification to obtain and to evaluate epidemiologic data on the prevalence of malocclusion in a group of 2,378 Yonsei University students, 17 to 23 years of age. All freshmen were selected, except for those students receiving orthodontic treatment and those few with too many missing teeth which prohibits classification by Angle's method. The following results were obtained: 1) Almost $91\%$ of students had malocclusion of the teeth severe enough to require correction. 2) There was a statistically significant difference in malocclusion between males and females($93.66\%$ malocclusion in males, $79.13\%$ malocclusioa in females). 3) Crowding was most pravalent in class I malocclusion. 4) There appeared to be a specific association between the number of lost first molars and Angle's classification. 5) In this study, more class II, Div.2 malocclusion appeared than in Massier's and Frankel's study of Caucasians, which used similar criteria. Class III malocclusion was more prevalent than normal occlusion in the Korean students studied, but in Caucasians' normal occlusion was more prevalent.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제11권5호
• /
• pp.2523-2538
• /
• 2017

In most of the existing literature, the definition of the class label has the following characteristics. First, the class label of the samples from the same object has an absolutely fixed value. Second, the difference between class labels of the samples from different objects should be maximized. However, the appearance of a face varies greatly due to the variations of the illumination, pose, and expression. Therefore, the previous definition of class label is not quite reasonable. Inspired by discriminative least squares regression algorithm (DLSR), a noisy label based discriminative least squares regression algorithm (NLDLSR) is presented in this paper. In our algorithm, the maximization difference between the class labels of the samples from different objects should be satisfied. Meanwhile, the class label of the different samples from the same object is allowed to have small difference, which is consistent with the fact that the different samples from the same object have some differences. In addition, the proposed NLDLSR is expanded to the kernel space, and we further propose a novel kernel noisy label based discriminative least squares regression algorithm (KNLDLSR). A large number of experiments show that our proposed algorithms can achieve very good performance.

• Journal of the Korean Data and Information Science Society
• /
• 제17권2호
• /
• pp.449-453
• /
• 2006

In teaching statistics in the secondary school, most math teachers do it and students enjoy it well before 11th grade. But, both of teachers and students feel statistics hard at 11th grade, which has been contained permutations, combinations, random variables, probability distributions, hypotheses and tests for normal distribution. In this study, we explore the efficient teaching methods of statistics for math teachers and also find many students enjoying it.

• 대한수학회지
• /
• 제53권1호
• /
• pp.233-246
• /
• 2016

In this paper, we introduce the class of analytic extensions of M-hyponormal operators and we study various properties of this class. We also use a special Sobolev space to show that every analytic extension of an M-hyponormal operator T is subscalar of order 2k + 2. Finally we obtain that an analytic extension of an M-hyponormal operator satisfies Weyl's theorem.

• Imaging Science in Dentistry
• /
• 제48권1호
• /
• pp.51-57
• /
• 2018

Purpose: To perform a comparative analysis of the palatal bone thickness in Thai patients exhibiting class I malocclusion according to whether they exhibited a normal or open vertical skeletal configuration using cone-beam computed tomography (CBCT). Materials and Methods: Thirty CBCT images of Thai orthodontic patients (15-30 years of age) exhibiting class I malocclusion with a normal or open vertical skeletal configuration were selected. Palatal bone thickness was measured in a 3.0-mm grid pattern on both the right and left sides. The palatal bone thickness of the normal-bite and open-bite groups was compared using the independent t-test. The level of significance was established at P<.05. Results: The palatal bone thickness in the normal-bite group ranged from $2.2{\pm}1.0mm$ to $12.6{\pm}4.1mm$. The palatal bone thickness in the open-bite group ranged from $1.9{\pm}1.1mm$ to $13.2{\pm}2.3mm$. The palatal bone thickness was lower at almost all sites in patients with open bite than in those with normal bite. Significant differences were found at almost all anteroposterior sites along the 3 most medial sections (3.0, 6.0, and 9.0 mm lateral to the midsagittal plane)(P<.05). Conclusion: Class I malocclusion with open vertical skeletal configuration may affect palatal bone thickness, so the placement of temporary anchorage devices or miniscrew implants in the palatal area in such patients should be performed with caution.

• 대한지구과학교육학회지
• /
• 제7권1호
• /
• pp.133-143
• /
• 2014

The purpose of this study was to examine the effect of scientific discussion classes focusing problem finding on the primary school students' scientific creative problem solving ability, science process skills and attitude toward science class. To verify this research problem, the subject of this study was fifth-grade students selected from four classes of M elementary school located in Busan city. For four months, the experimental group of 51 students was taught using the "scientific discussion classes focusing problem finding". The control group also of 53 students was taught in normal classes which used a text-book. All students were given pre and post test to verify the effects of scientific discussion classes focusing problem finding on the primary school students' scientific creative problem solving ability, science process skills and attitude toward science class. The results from this study are as the following. First, the scientific discussion classes focusing problem finding were effective in scientific creative problem solving ability among the primary school students. It is possibly because in the process where one student compare his/her own thoughts with the others' ones and discuss them. Second, the scientific discussion classes focusing problem finding were effective in science process skills among the primary school students. Third, the scientific discussion classes focusing problem finding were effective in attitude toward science class. In conclusion, the scientific discussion classes focusing problem finding had positive effects on improvement of primary school students' scientific creative problem solving ability, science process skills and also could lead to a change in students' cognition about science class to a positive way. Therefore, the scientific discussion class focusing problem finding is hopefully to be provided as an effective instructive strategy of science class in school in the future.

• 대한수학회보
• /
• 제45권1호
• /
• pp.23-31
• /
• 2008

In this paper, we investigate nonoscillatory solutions of a class of higher order neutral nonlinear difference equations with positive and negative coefficients $\Delta^m(x(n)+p(n)x(\tau(n)))+f_1(n,x(\sigma_1(n)))-f_2(n,x(\sigma_2(n)))=0,\;n{\geq}n_0$. Some sufficient conditions for the existence of nonoscillatory solutions are obtained.

• 디지털융복합연구
• /
• 제14권12호
• /
• pp.607-614
• /
• 2016

본 연구는 체력관리 수업이 비만도가 각기 다른 여자 대학생들의 신체구성, 혈압 및 건강관련 체력에 미치는 영향이 다른지에 대해 검증하는 것이다. 본 연구의 대상자는 본교 2012~2016년 9학기 동안 체력관리 수업을 수강한 여학생 317명을 대상으로 하였으며, 그룹 구분은 체질량지수와 체지방률(%) 기준으로 저체중, 마른 비만, 정상, 정상 비만 그리고 비만 그룹으로 구분하였다. 체력관리 수업은 저항성 운동과 유산소 운동의 복합운동으로 일일 75분, 주 2회로 구성하여 진행하였다. 본교의 체력관리 수업은 비만 그룹을 제외한 네 그룹의 체중, 체질량지수, 체지방율, 근육량 그리고 허리둘레를 긍정적으로 개선시키는 효과가 있는 것으로 나타났다. 하지만 혈압(SBP, DBP)과 건강관련 체력에서는 다섯 그룹 모두에서 대부분 긍정적으로 개선되는 효과가 있었다. 따라서 체력관리 수업은 비만 여자 대학생의 신체구성의 개선을 제외하고 여자 대학생들의 신체구성과 혈압의 개선 및 건강관련 체력을 향상시키는 것으로 나타났다.

• Journal of applied mathematics & informatics
• /
• 제20권1_2호
• /
• pp.465-478
• /
• 2006

In this paper, we introduce a new class of completely generalized strongly nonlinear quasivariational inequalities and establish its equivalence with a class of fixed point problems by using the resolvent operator technique. Utilizing this equivalence, we develop a three-step iterative scheme with errors, obtain a few existence theorems of solutions for the completely generalized non-linear strongly quasivariational inequality involving relaxed monotone, relaxed Lipschitz, strongly monotone and generalized pseudocontractive mappings and prove some convergence and stability results of the sequence generated by the three-step iterative scheme with errors. Our results include several previously known results as special cases.