Earticle

본 연구는 초등학교 학생들을 대상으로 반응 중심 학습법을 통한 동기유발이 체육수업 전·후의 참여태도에 미치는 영향과 변화를 분석하여 체육수업의 실제를 파악할 수 있는 관점을 제공하고자 하는데 목적이 있다. 이를 위해 본 연구의 대상은 C지역 소재 N초등학교 5, 6학년 126명의 학생을 학급과 성별 구분 없이 선정하였으며, 탐색적 요인분석과 요인 간 유사성을 알아보기 위한 적률상관계수를 적용하여 분석하였다. 본 연구결과에서 제시된 것과 같이 일방적인 주입식 교육이 아닌 학생들을 중심으로 한 교육으로 추진해 나가는 것이 필요하며 학생들이 적극적이고 활발한 체육수업 참여에 기여할 수 있도록 수업 전·후의 학습요인을 고려하여 융합적으로 적용하는 것이 중요하다.

This study aims to analyze the impact of motivation through Response-Based Learning (RBL) on elementary students’ attitude of participation and change before and after a Physical Education class and provide a perspective on the understanding of the practice of the PE class. For this purpose, this study selected 115 students in fifth grade in N. Elementary School located in the C. region as subjects, regardless of their sex. An exploratory factorial analysis was conducted with the subjects, and they were analyzed, applying a product-moment coefficient to examine the similarity between the factors. As shown in the result of this study, it is necessary to promote student-centered education instead of one-sided cramming education. Also, it is important to consider and convergence on apply the factors of learning before and after the class to be able to contribute to students’ active and positive participation in the PE class.

간세포암 환자에서 TACE 치료 반응은 환자의 개별적인 특성에 따라 달라지므로, 치료 효과를 정확히 예측하는 것은 여전히 중요한 도전 과제로 남아 있다. 이에 본 연구에서는 다양한 머신러닝 알고리즘(Logistic Regression, Linear Support Vector Machine, Random Forest, XGBoost)을 활용하여 TACE 치료 반응을 예측하는 모델을 개발하고, 이들의 성능을 비교하여 가장 우수한 모델을 도출하고자 하였다. 나이, 성별, BMI, 간염 바이러스, 간경변증, 문맥 고혈 압, 생체표지자, CBCT, DAP을 독립변수로, 치료 완전반응을 종속변수로 설정하여 예측 모델을 개발하였다. 성능 평가 결과, 선형 모델 중에서는 Logistic Regression이 가장 높은 성능을 보였으며, 비선형 모델에서는 Random Forest가 재현율(82.12%)에서 높은 값을 나타냈다. 그러나 XGBoost는 정확도(78.57%), 정밀도(81.43%), F1 점수(81.71%)에서 더 높은 성능을 보여 종합적으로 가장 우수한 예측력을 보였다. 이러한 결과는 XGBoost 모델이 복잡한 임상 데이터를 효과적으로 처리할 수 있으며, 향후 TACE 치료 계획 수립에 유용하게 활용될 것으로 사료된다.

Treatment response to TACE (Transarterial chemoembolization) in patients with HCC (Hepatocellular carcinoma) varies depending on individual patient variables, making accurate prediction of treatment outcomes a significant challenge. Therefore, this study aimed to develop Prediction models for treatment response using various machine learning algorithms, including Logistic Regression, Linear Support Vector Machine, Random Forest, and XGBoost, and to compare their performance to identify the most effective model. The Prediction models were developed using independent variables such as age, sex, BMI (Body Mass Index), hepatitis virus, liver cirrhosis, portal hypertension, biomarker, CBCT (Cone Beam Computed Tomography), and DAP (Dose Area Product), with complete treatment response set as the dependent variable. Performance evaluation showed that Logistic Regression had the highest performance among linear models, while Random Forest demonstrated superior recall (82.12%) among nonlinear models. However, XGBoost outperformed other models in terms of accuracy (78.57%), precision (81.43%), and F1 score (81.71%), demonstrating the best overall predictive performance. These results suggest that the XGBoost model can effectively handle complex clinical data and may be useful in supporting future TACE treatment planning.

최근 컴퓨터 비전 모델의 성능이 급속도로 향상되고 있다. 하지만 모델의 파라미터 수가 증가하여 모바일 환경에서는 자원의 제약으로 모델 크기와 연산 시간을 최적화하는 경량화 연구가 필요하다. 경 량화를 위한 효과적인 방법 중 하나인 지식 증류는 잘못된 확률적 지식까지 증류하는 문제를 가진다. 본 논문은 클래스별 적응형 온도 지식 증류를 통해 잘못된 확률적 지식을 증류하는 문제를 해결하고, 더 정제된 지식을 학생 모델에 전달하는 방법을 제안한다. 실험 결과, 제안한 방법이 기존 지식 증류 기법들보다 우수한 성능을 보임을 알 수 있다.

쿠버네티스는 컨테이너 통합 관리를 위한 대표적인 오픈소스 기반 소프트웨어로, 컨테이너에 할당된 자원을 모니터링하고 관리하는 핵심적인 역할을 한다. 컨테이너 환경이 보편화됨에 따라 컨테이너를 대상으로 한 보안 위협이 지속적으로 증가하고 있으며, 대표적인 공격으로는 자원 고갈 공격이 있다. 이는 악성 크립토마이닝 소프트웨어를 컨테이너 형태로 배포하여 자원 을 탈취함으로써, 자원을 공유하는 호스트 및 다른 컨테이너의 동작에 영향을 끼친다. 선행 연구는 자원 고갈 공격의 탐지에 초점이 맞춰져 있어 공격 발생 시 대응하는 기술은 부족한 실정이다. 본 논문은 쿠버네티스 환경에서 구동되는 컨테이너를 대 상으로 한 자원 고갈 공격 및 악성 컨테이너를 탐지하고 대응하기 위한 강화학습 기반 동적 자원 관리 프레임워크를 제안한 다. 이를 위해, 자원 고갈 공격 대응 관점에서의 강화학습 적용을 위한 환경의 상태, 행동, 보상을 정의하였다. 제안한 방법론 을 통해, 컨테이너 환경에서의 자원 고갈 공격에 강인한 환경을 구축하는 데 기여할 것으로 기대한다.

Kubernetes is a representative open-source software for container orchestration, playing a crucial role in monitoring a nd managing resources allocated to containers. As container environments become prevalent, security threats targeting con tainers continue to rise, with resource exhaustion attacks being a prominent example. These attacks involve distributing m alicious crypto-mining software in containerized form to hijack computing resources, thereby affecting the operation of the host and other containers that share resources. Previous research has focused on detecting resource depletion attacks, so t echnology to respond when attacks occur is lacking. This paper proposes a reinforcement learning-based dynamic resource management framework for detecting and responding to resource exhaustion attacks and malicious containers running in Kubernetes environments. To achieve this, we define the environment's state, actions, and rewards from the perspective of responding to resource exhaustion attacks using reinforcement learning. It is expected that the proposed methodology will contribute to establishing a robust defense against resource exhaustion attacks in container environments

최근 기술 발달에 따라 산업기술 유출 사고를 막기 위한 다양한 보안대책과 보안 시설 및 시 스템도 고도화 되었다. 그러나 이와 같은 노력에도 산업기술 유출 사고는 지속적으로 발생하고 있어 산업기술 유출 사고 예방은 상당히 어려운 과제라는 인식을 주었다. 산업기술 유출 사고는 내부적으로만 발생하는 것도 아니고, 외부적으로만 발생하는 것도 아니 다. 따라서 사용자 행위 정보를 수집 한다고, 쉽게 산업기술 유출 행위를 식별하거나 특정하기 어려운 점이 있다. 그렇기 때문에 선행연구 자료들에는 다양한 관점에서의 산업기술 유출 행위 를 분석하였고, 해당 행위를 탐지하기 위한 항목들을 도출하였다. 하지만 이러한 항목들은 사용 자의 움직임 행위가 아닌 온라인상의 행동 정보를 기반으로 한정되어 있다. 산업기술 유출 대응에 가장 많이 사용되고 있는 것은 보안 솔루션들이나, 이 보안 솔루션들은 사용자의 움직임 행위 즉 행동 인식이 아닌 개인 또는 조직에 대한 온라인상의 로그를 기반으로 행동을 유추하여 분석하는 특징을 가지고 있다. 따라서 이 보안 솔루션은 DB 또는 PC에 저장되 어 있는 내부 정보를 외부로 전송하거나, 접근 또는 출력하는 것에 대한 통제를 담당하는 것이기 때문에 직접적인 사용자의 행동 인지는 하지 못한다는 문제점이 있다. 또 기술적인 부분이 아닌 정책인 부분은 사실상 산업기술 유출 사고가 발생한 후에 적용되는 것이므로, 사전에 산업기술 유출을 방지하기 위해서는 온라인상의 행동 정보를 분석하는 보안 솔루션에 추가적인 요소가 필요하다. 그 추가적인 요소는 바로 사용자의 행동을 학습하는 것이며, 사용자의 행동 인식 학습 정보와 기존 보안솔루션의 학습 정보를 통합 분석한다면, 산업기술 유출 사고에 대한 예방적 보안대책 이 될 수 있다고 판단하여 WIFI 신호 활용 행동 인식 학습 기반 산업기술 유출 대응 시스템 설계 제안하고자 한다.

In accordance with recent technological development, various security measures and security facilities and systems have been advanced to prevent leakage of industrial technology. However, despite these efforts, industrial technology leakage accidents continue to occur, giving recognition that preventing industrial technology leakage accidents is a very difficult task. Industrial technology leakage accidents do not only occur internally, nor do they only occur externally. Therefore, even if user behavior information is collected, it is difficult to easily identify or specify the act of leaking industrial technology. Therefore, in the previous research data, the industrial technology leakage behavior was analyzed from various perspectives, and items for detecting the corresponding behavior were derived. However, these items are limited based on online behavioral information rather than the user's movement behavior. Security solutions are most often used to respond to industrial technology leaks, but these security solutions have the characteristic of inferring and analyzing behavior based on online logs for individuals or organizations, rather than recognizing user's movement behavior, that is, behavior. Therefore, since this security solution is in charge of controlling external transmission, access, or output of internal information stored in a DB or PC, there is a problem in that it does not directly recognize the user's actions. In addition, since the policy part, not the technical part, is actually applied after an industrial technology leakage accident occurs, additional elements are needed in the security solution that analyzes online behavioral information to prevent industrial technology leakage in advance. The additional element is to learn the user's behavior, and if the user's behavior recognition learning information and the learning information of the existing security solution are integrated and analyzed, it can be a preventive security measure against industrial technology leakage accidents, and the WIFI signal is used. The design of an industrial technology leak response system based on behavioral recognition learning is proposed.

본 연구는 게임기반 학생응답시스템을 활용한 수업이 학습몰입에 미치는 영향을 알아보는 데 그 목적이 있다. 간호교육에서 게임기반 학생응답시스템을 활용한 수업(실험집단)과 강의식 수업(비교집단)의 학습몰입을 중심으로 비교 한 실험연구이다. 2019년 10월부터 2019년 12월까지 일개 간호대학생 211명이 참여하였다. 학습몰입의 두 집단 간의 차이는 t-test를, 관련 요인과는 상관분석을 시행하였다. 비교집단과 실험집단 간 학습몰입의 차이를 분석한 결과, 전체 학습몰입(p=.013), 정서몰입(p=.002)에서 차이를 보였다. 이는 국내 처음으로 게임기반 학생응답시스템을 활용한 수업 이 학습몰입의 효과성을 검증했다는 점에서 그 의의가 있다.

The purpose of this study is to find out the impact of classes using a game-based student response system on learning engagement. It is an experimental study that compares learning engagement in classes (experimental groups) and lecture-style classes (comparative groups) that utilize GSRS in nursing education. A total of 211 nursing students participated from October 2019 to December 2019. The differences in learning engagement between the two groups were analyzed as t-test and correlation analysis was conducted on related factors. There was a difference between the comparison group and the experimental group in overall learning engagement(p=.013) and emotional engagement(p=.002). This is meaningful in that it has verified the learning engagement effect of the GSRS for the first time in Korea.

화학공장의 누출사고는 초기에 적절히 대응하지 못할 경우 화재·폭발과 같은 2차·3차의 복합재난사고로 확산될 위험성이 매우 높다. 이러한 이유로 누출사고 발생 초기에 누출이 발생한 지점을 신속히 파악하여 현장안전요원에게 알림으로써, 보다 체계적이고 효율적인 초기대응을 가능하게 하여, 사고피해를 완화시킬 수 있는 통합적인 누출사고 대응시스템 구축은 매우 중요하다고 할 수 있다. 본 연구에서는, 통합적인 누 출사고 대응시스템 구축을 위한 선행연구로, 딥러닝 기반의 누출원추적 모델 개발을 제안한다. 여수에 위 치한 실제 화학공장을 대상으로 누출사고 시나리오에 대한 Computational Fluid Dynamics (CFD) 시뮬레 이션을 진행한 뒤, 화학공장 경계면에 배치된 각 센서별 위치에서의 농도, 풍향 그리고 풍속데이터를 추출 하고, 센서 좌표를 추가하여 인공신경망을 학습시켰다. 학습된 모델은 40개의 누출후보군에 대해 학습에 사용되지 않은 상황들에서도 75.43%의 정확도로 누출이 일어난 지점을 실시간 예측해냄을 확인하였다. 또 한 누출지점 예측이 일치하지 않은 경우도, 예측된 지점이 실제 누출이 일어난 지점과 물리적으로 매우 인 접함을 확인함으로써 제안된 모델을 실제 현장에 적용할시 기대되는 효과는 더 클 것으로 판단하였다.

본 연구의 목적은 시뮬레이션 기반 흉관배액관리 간호교육이 간호학생의 시나리오 경험에 대한 반응, 학습에 대한 자신감 및 문제해결능력에 미치는 효과를 검증하는 것이다, 본 연구는 단일군 전후설계 실험 연구로 133명의 3학 년 간호학생을 31개조로 시뮬레이션 교육 중재를 시행하였다. 연구결과, 시뮬레이션 실습교육 시행 전보다 시행 후에 시나리오 경험에 대한 반응은 긍정적인 결과로 나타났고, 학습에 대한 자신감 및 문제해결능력은 유의하게 향상된 것으 로 나타났다. 이는 간호학 실습교육에 있어서 시뮬레이터를 이용한 시뮬레이션 교육이 효과적임을 보여주는 결과라고 할 수 있다. 본 연구는 현장중심의 실습교육 및 임상실무의 질을 향상시키는데 기여할 수 있을 것이다.

The purpose of this study is to verify the effects of Simulation-based Chest Tube Drainage Management Nursing Education on the Nursing Students’ Response to Scenario Experiences, Confidence in Learning, and Problem Solving Abilities. This study was performed with a one-group pretest-posttest design, and a total of 133 junior nursing students have participated in the research (31 teams of simulation-based education). According to the results of the study, most of the students have shown a positive response to the scenario experience after the simulation training, while their confidence in learning and problem-solving skills have improved significantly. This proves that simulation-based training using simulators can be very effective in practical nursing training. It is expected that this research could make a meaningful contribution to improving the overall quality of field-oriented practical training and clinical practice.

In this paper, a Price learning based Load Distribution Strategy (PLDS) is proposed at first. In PLDS model, Smart Power Service, Utility Company and History Load Curves are included, and by considering both the average electricity consumption cost and the average electricity consumption habit, we proposed a convex optimization model to solve the model. In order to accelerate the convergence of PLDS, a price learning mechanism is proposed, which learns a price curve according to the history price data, and predicts price as a learned price for the next iteration. The optimization cycle of PLDS is one day or 24 hours, and in order to further improve the peak shaving performance, an extended version of PLDS named PLRS (Price learning based Load Redistribution Strategy) is proposed, whose optimization cycle length is 1 hour. The optimization models of PLDS and PLRS are the same, and the differences between them are the optimization cycle and the constraint conditions. In the simulation, we compared the convergence performance, peaking shaving performance and total cost among PLDS, PLRS and other strategy ODC in reference [11], and we found that the convergence performances of PLDS and PLRS are both better than that of ODC. The peak shaving performance of PLRS is better than that of ODC in the long term, and the total cost of PLRS is very close to that of ODC.

Recently, as utilization rate of the Internet gets higher, cybercrime rates are growing up. In order not to be damaged from such cybercrimes, the importance of an education for protection from them is increasingly critical. However, the existing information security educations have not been enough to deal effectively with rising cybercrimes. To tackle the problems of the existing educations, in this paper, we examine a way to prevent cybercrimes using learning contents related to a real life, help learners to improve awareness of cybercrimes and realize their seriousness through an indirect experience of them. We develop a program with learning contents based on a simulation for preventing cybercrimes. Furthermore, we provide a quiz game to enhance the understanding of the cybercrimes. After learning through the quiz game, the learners can check and extend their knowledge of preventing the cybercrimes. Specifically, each learner can identify vulnerable cybercrime types and get the proper feedbacks for preventing them through the analysis of learner’s own correct/wrong answers.

이 논문에서 우리는 응급상황에 대응하여 일정 구간의 교통신호를 능동적으로 제어함으로써 재산과 인명 손실을 최소화할 수 있는 응급상황 대응 교통신호 제어 시스템을 개발하였다. 응급 차량 단말기에서 식별정보 및 GPS 정보를 포함한 응급신호를 송출하면 카메라에서 주위 영상을 획득하게 되고, 딥러닝 기반으로 객체를 분석하여 객체의 위치, 종류, 크기 등 정보를 가지는 객체정보를 출력한다. 이 객체를 트래킹한 정보를 생성하여 신호체계를 검출한 후 신호체계를 응급모드로 전환하여 수신받은 GPS 정보를 기준으로 응급 차량을 식별·추적하고 이 응급 차량의 진행 경로 기준으로 긴급 제어신호를 교통신호 제어기로 전송할 수 있는 체계이다. 이 시스템은 응급신호에 따라 우선 적용되는 긴급 제어신호에 의해 응급 차량의 진행이 저지되지 않도록 하여, 교통상 장애에 따른 인명과 재산의 손실을 최소화할 수 있다.

In this paper, we developed a traffic signal control system for emergency situations that can minimize loss of property and life by actively controlling traffic signals in a certain section in response to emergency situations. When the emergency vehicle terminal transmits an emergency signal including identification information and GPS information, the surrounding image is obtained from the camera, and the object is analyzed based on deep learning to output object information having information such as the location, type, and size of the object. After generating information tracking this object and detecting the signal system, the signal system is switched to emergency mode to identify and track the emergency vehicle based on the received GPS information, and to transmit emergency control signals based on the emergency vehicle's traveling route. It is a system that can be transmitted to a signal controller. This system prevents the emergency vehicle from being blocked by an emergency control signal that is applied first according to an emergency signal, thereby minimizing loss of life and property due to traffic obstacles.

Although many models have been proposed to accurately predict the response of drugs in cell lines recent years, understanding the genome related to drug response is also the key for completing oncology precision medicine. In this paper, based on the cancer cell line gene expression and the drug response data, we established a reliable and accurate drug response prediction model and found predictor genes for some drugs of interest. To this end, we first performed pre-selection of genes based on the Pearson correlation coefficient and then used ElasticNet regression model for drug response prediction and fine gene selection. To find more reliable set of predictor genes, we performed regression twice for each drug, one with IC₅₀ and the other with area under the curve (AUC) (or activity area). For the 12 drugs we tested, the predictive performance in terms of Pearson correlation coefficient exceeded 0.6 and the highest one was 17-AAG for which Pearson correlation coefficient was 0.811 for IC₅₀ and 0.81 for AUC. We identify common predictor genes for IC₅₀ and AUC, with which the performance was similar to those with genes separately found for IC₅₀ and AUC, but with much smaller number of predictor genes. By using only common predictor genes, the highest performance was AZD6244 (0.8016 for IC₅₀, 0.7945 for AUC) with 321 predictor genes.

Background: Hepatocellular carcinoma (HCC) remains a leading cause of cancer-related mortality worldwide, largely because of challenges in early diagnosis and the limited sensitivity of conventional biomarkers. Therefore, reliable molecular tools for early detection, prognostic stratification, and individualized treatment predictions are urgently required.Methods: This retrospective study analyzed publicly available gene expression datasets. Candidate biomarkers were identified from the GSE14520 cohort using a multistep screening workflow that integrated differential expression analysis, diagnostic performance, and prognostic relevance. A 10-gene diagnostic model was constructed using least absolute shrinkage and selection operator logistic regression and subsequently validated across multiple independent cohorts. Survival outcomes were evaluated using the Kaplan-Meier analysis and treatment responses to sorafenib and transarterial chemoembolization (TACE) were assessed using receiver operating characteristic analysis.Results: A 10-gene signature (<i>TOP2A, CDK1, CYP3A4, MASP2, EPHX2, HAO1, RACGAP1, GLYAT, ADH1B</i>, and <i>CYP4A11</i>) was established. The model demonstrated robust internal performance and consistent accuracy across external validation cohorts (area under the curve [AUC], >0.9). This signature effectively identified early-stage HCC and distinguished malignancy from cirrhosis. High-risk scores were significantly associated with poor overall survival and recurrence-free survival (<i>p</i><0.05). Furthermore, the model could predict treatment sensitivity, with higher risk scores associated with better outcomes for sorafenib (AUC, 0.791), whereas lower risk scores correlated with an improved response to TACE (AUC, 0.768).Conclusion: Our gene expression-based machine learning model provides a robust tool for HCC diagnosis, prognosis, and treatment response prediction, with potential as a supportive system for personalized clinical decision-making.

Second language (L2) peer response literature is defined in part by discourse research, yet there is scant research on text-specific comments, or comments that make explicit text references, thus resisting generic qualities. The purpose of this case study was to examine such peer response activities in an English writing course at a South Korean university. The data comprises two essay assignments with peer response conducted between two drafts ? as accomplished during class time on the class learning management system (LMS) ? as well as the subsequent revisions in second drafts. This paper expands on previous coding schemes accounting for area, nature and type commentary to account for a specificity dimension, and also links these categories to revision practices. While students entertained diverse commenting and revising options, popular practices included generic evaluating or revising local or surface-level concerns. This paper offers implications for modelling response activities as well as for how to better define specific and complex idea construction exhibited during response.

In this paper, machine learning models were applied to predict the seismic response of steel frame structures. Both geometric and material nonlinearities were considered in the structural analysis, and nonlinear inelastic dynamic analysis was performed. The ground acceleration response of the El Centro earthquake was applied to obtain the displacement of the top floor, which was used as the dataset for the machine learning methods. Learning was performed using two methods: Decision Tree and Random Forest, and their efficiency was demonstrated through application to 2-story and 6-story 3-D steel frame structure examples.

Machine learning is widely applied to various engineering fields. In structural engineering area, machine learning is generally used to predict structural responses of building structures. The aging deterioration of reinforced concrete structure affects its structural behavior. Therefore, the aging deterioration of R.C. structure should be consider to exactly predict seismic responses of the structure. In this study, the machine learning based seismic response prediction model was developed. To this end, four machine learning algorithms were employed and prediction performance of each algorithm was compared. A 3-story coupled shear wall structure was selected as an example structure for numerical simulation. Artificial ground motions were generated based on domestic site characteristics. Elastic modulus, damping ratio and density were changed to considering concrete degradation due to chloride penetration and carbonation, etc. Various intensity measures were used input parameters of the training database. Performance evaluation was performed using metrics like root mean square error, mean square error, mean absolute error, and coefficient of determination. The optimization of hyperparameters was achieved through k-fold cross-validation and grid search techniques. The analysis results show that neural networks and extreme gradient boosting algorithms present good prediction performance.

국가 지원 사이버 공격은 사전에 계획된 목표를 달성하기 위하여 수행되기 때문에 그 파급력이 크다. 방어자 입장에서 이에 대응을 해야하지만 공격의 규모가 크고 알려지지 않은 취약점이 활용될 가능성도 있기 때문에 대응하기 어렵다. 또한 너무 과한 대응은 사용자의 업무의 가용성을 떨어뜨려서 업무에 지장이 생길 수 있다. 따라서 사용자의 가용성을 확보하면서도 효율적으로 공격을 방어할 수 있는 대응 정책이 필요하다. 본 논문에서는 이를 해결하기 위하여 실시간으로 방어 자산의 프로세스 수와 세션 수를 수집하여 학습에 활용하는 방법을 제안한다. 해당 방법을 활용하여 사이버 공격 시뮬레이터 상에서 강화학습 기반 정책을 학습한 결과, 두 가지 공격자 모델에 대하여 100 time-steps 기준 공격 지속 시간은 각 27.9 time-steps, 3.1 time-steps만큼 감소시켰으며 또한 방어 과정에서 사용자의 가용성을 저해시키는 "복원"행위의 횟수도 감소하여 종합적으로 더 좋은 성능의 정책을 도출할 수 있었다.

State-sponsored cyber attacks are highly impactful because they are carried out to achieve pre-planned goals. As a defender, it is difficult to respond to them because of the large scale of the attack and the possibility that unknown vulnerabilities may be exploited. In addition, overreacting can reduce the availability of users and cause business disruption. Therefore, there is a need for a response policy that can effectively defend against attacks while ensuring user availability. To solve this problem, this paper proposes a method to collect the number of processes and sessions of defense assets in real time and use them for learning. Using this method to learn reinforcement learning-based policies on a cyber attack simulator, the attack duration based on 100 time-steps was reduced by 27.9 time-steps and 3.1 time-steps for two attacker models, respectively, and the number of "restore" actions that impede user availability during the defense process was also reduced, resulting in an overall better policy.