• Journal of Korea Port Economic Association
• /
• v.39 no.2
• /
• pp.59-76
• /
• 2023

In this study, we tried to find out port safety from various perspectives through news data that can be easily accessed by the general public and domestic academic journal data that reflects the insights of port researchers. Non-negative Matrix Factorization(NMF) based topic modeling was conducted using Python to derive the main topics for each data, and then semantic analysis was conducted for each topic. The news data mainly derived natural and environmental factors among port safety risk factors, and the academic journal data derived security factors, mechanical factors, human factors, environmental factors, and natural factors. Through this, the need for strategies to strengthen the safety of domestic ports, such as strengthening the resilience of port safety, improve safety awareness to broaden the public's view of port safety, and conduct research to develop the port industry environment into a safe and specialized mature port. As a result, this study identified the main factors to be improved and provided basic data to develop into a mature port with a port safety culture.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.10
• /
• pp.67-76
• /
• 2023

Although deep learning models are making innovative achievements in the field of computer vision, the problem of vulnerability to adversarial examples continues to be raised. Adversarial examples are attack methods that inject fine noise into images to induce misclassification, which can pose a serious threat to the application of deep learning models in the real world. In this paper, we propose a model that detects adversarial examples using differences in predictive values between edge-learned classification models and underlying classification models. The simple process of extracting the edges of the objects and reflecting them in learning can increase the robustness of the classification model, and economical and efficient detection is possible by detecting adversarial examples through differences in predictions between models. In our experiments, the general model showed accuracy of {49.9%, 29.84%, 18.46%, 4.95%, 3.36%} for adversarial examples (eps={0.02, 0.05, 0.1, 0.2, 0.3}), whereas the Canny edge model showed accuracy of {82.58%, 65.96%, 46.71%, 24.94%, 13.41%} and other edge models showed a similar level of accuracy also, indicating that the edge model was more robust against adversarial examples. In addition, adversarial example detection using differences in predictions between models revealed detection rates of {85.47%, 84.64%, 91.44%, 95.47%, and 87.61%} for each epsilon-specific adversarial example. It is expected that this study will contribute to improving the reliability of deep learning models in related research and application industries such as medical, autonomous driving, security, and national defense.

• Smart Media Journal
• /
• v.13 no.6
• /
• pp.9-15
• /
• 2024

In recent trends, there has been an increase in 'Qshing' attacks, a hybrid form of phishing that exploits fake QR (Quick Response) codes impersonating government agencies to steal personal and financial information. Particularly, this attack method is characterized by its stealthiness, as victims can be redirected to phishing pages or led to download malicious software simply by scanning a QR code, making it difficult for them to realize they have been targeted. In this paper, we have developed a classification technique utilizing machine learning algorithms to identify the maliciousness of URLs embedded in QR codes, and we have explored ways to integrate this with existing QR code readers. To this end, we constructed a dataset from 128,587 malicious URLs and 428,102 benign URLs, extracting 35 different features such as protocol and parameters, and used AutoML to identify the optimal algorithm and hyperparameters, achieving an accuracy of approximately 87.37%. Following this, we designed the integration of the trained classification model with existing QR code readers to implement a service capable of countering Qshing attacks. In conclusion, our findings confirm that deriving an optimized algorithm for classifying malicious URLs in QR codes and integrating it with existing QR code readers presents a viable solution to combat Qshing attacks.

• Industry Promotion Research
• /
• v.9 no.1
• /
• pp.123-132
• /
• 2024

The purpose of this study was to improve the understanding of stress coping in the very old and elderly and to improve life satisfaction through stress reduction as the population ages. To this end, this study aimed to test the mediating effect of social support on the relationship between stress coping and life satisfaction among very old people. In particular, we paid attention to the differences in stress coping from the existing elderly population and took the view that they should be studied as an independent group. To confirm this, we analysed 275 men and women aged 85 years and older who responded to the fifth supplementary survey of the National Elderly Security Panel (KReIS) conducted by the National Pension Service. IBM SPSS 26 was used to test the mediating effect of social support on the effect of stress coping measures on life satisfaction in the very old elderly. The results of the significance test of the independent variables on the mediating variable showed that stress coping was positively significant, i.e., the higher the level of stress coping, the higher the life satisfaction. In addition, the results of the significance test of the effects of the independent variables and mediators on the dependent variable showed that coping with stress had a significant effect, and the mediator, social support, also had a significant effect on life satisfaction. Therefore, this study suggests the need for social support to improve the level of life satisfaction through coping with stress in the very old elderly.

• Information Systems Review
• /
• v.20 no.4
• /
• pp.43-57
• /
• 2018

We propose network effects upon the investment decision of cloud-based ERP. Using the survey data collected from 82 companies in 2015, we examine whether IT managers have an intention to adopt real options in order to manage the risk of cloud-based ERP investments and how the network effects influence upon the intention to adopt real options. Based on prior literature, we propose a research model with 4 hypotheses. We find partial support of the hypotheses from the empirical analysis: technological risks has a positive impact upon the adoption of real options such as defer, contract, and abandon. In contrast, we find no significant impact of security risks upon real options. We validate positive network effects upon the adoption of real options such as defer, contract, and abandon. This work empirically find that IT managers in Korean middle and small sized firms have an intention to adopt real options when the managers realize economic, technological, and relationship risks and when they expect network effects.

• Journal of the Korean Society for information Management
• /
• v.41 no.2
• /
• pp.245-268
• /
• 2024

As the recognition of data's value increases, the role of data repositories in managing, preserving, and utilizing data is becoming increasingly important. This study investigates ways to enhance the trustworthiness of data repositories through obtaining CoreTrustSeal (CTS) certification. Trust in data repositories is critical not only for data protection but also for building and maintaining trust between the repository and stakeholders, which in turn affects researchers' decisions on depositing and utilizing data. The study examines the CoreTrustSeal, an international certification for trustworthy data repositories, analyzing its impact on the trustworthiness and efficiency of repositories. Using the example of DataON, Korea's first CTS-certified repository operated by the Korea Institute of Science and Technology Information (KISTI), the study compares and analyzes four repositories that have obtained CTS certification. These include DataON, the Physical Oceanography Distributed Active Archive Center (PO.DAAC) from NASA, Yareta from the University of Geneva, and the DARIAH-DE repository from Germany. The research assesses how these repositories meet the mandatory requirements set by CTS and proposes strategies for improving the trustworthiness of data repositories. Key findings indicate that obtaining CTS certification involves rigorous evaluation of organizational infrastructure, digital object management, and technological aspects. The study highlights the importance of transparent data processes, robust data quality assurance, enhanced accessibility and usability, sustainability, security measures, and compliance with legal and ethical standards. By implementing these strategies, data repositories can enhance their reliability and efficiency, ultimately promoting wider data sharing and utilization in the scientific community.

• The Journal of the Convergence on Culture Technology
• /
• v.10 no.2
• /
• pp.167-172
• /
• 2024

In major wars of the 21st century, drones are expanding beyond surveillance and reconnaissance to include land and air as well as sea and underwater for purposes such as precision strikes, suicide attacks, and cognitive warfare. These drones will perform multi-domain operations, and to this end, they will continue to develop by improving the level of autonomy and strengthening scalability based on the High-Low Mix concept. Recently, drones have been used as a major means in major wars around the world, and there seems to be a good chance that they will evolve into game changers in the future. North Korea has also been making significant efforts to operate reconnaissance and attack drones for a long time. North Korea has recently continued to engage in provocations using drones, and its capabilities are gradually becoming more sophisticated. In addition, with the recent emergence of new strategic Drones, wartime and peacetime threats such as North Korea's use of these to secure surveillance, reconnaissance and early warning capabilities against South Korea and new types of provocations are expected to be strengthened. Through this study, we hope to provide implications by analyzing the capabilities of North Korea's strategic Drones, predicting their operation patterns, and conducting active follow-up research on the establishment of a comprehensive strategy, such as our military's drone deployment and counter-drone system solutions.

• The Journal of the Convergence on Culture Technology
• /
• v.10 no.1
• /
• pp.183-192
• /
• 2024

The purpose of this study is to empirically analyze the determining factors (economy, security, domestic politics, administration, and international politics) that affect the ROK-US defense cost sharing decision. Through this, we will gain a deeper understanding of the defense cost sharing decision process and improve the efficiency of defense cost sharing calculation and execution. The scope of the study is ROK-US defense cost sharing from 1991 to 2021. The data used in the empirical analysis were various secondary data such as Ministry of National Defense, government statistical data, SIPRI, and media reports. As an empirical analysis method, multiple regression analysis using time series was used and the data was analyzed using an autoregressive model. As a result of empirical research through multiple regression analysis, we derived the following results. It was analyzed that the size of Korea's economy, that is, GDP, the previous year's defense cost share, and the number of U.S. troops stationed in Korea had a positive influence on the decision on defense cost sharing. This indicates that Korea's economic growth is a major factor influencing the increase in defense cost sharing, and that the gradual increase in the budget and the negotiation method of the Special Agreement (SMA) for cost sharing of stationing US troops in Korea play an important role. On the other hand, the political tendencies of the ruling party, North Korea's military threats, and China's defense budget were found to have no statistically significant influence on the decision to share defense costs.

• The Journal of Bigdata
• /
• v.8 no.2
• /
• pp.125-131
• /
• 2023

Globally, drug side effects rank among the top causes of death. To effectively respond to these adverse drug reactions, a shift towards an active real-time monitoring system, along with the standardization and quality improvement of data, is necessary. Integrating individual institutional data and utilizing large-scale data to enhance the accuracy of drug side effect predictions is critical. However, data sharing between institutions poses privacy concerns and involves varying data standards. To address this issue, our research adopts a federated learning approach, where data is not shared directly in compliance with privacy regulations, but rather the results of the model's learning are shared. We employ the Common Data Model (CDM) to standardize different data formats, ensuring accuracy and consistency of data. Additionally, we propose a drug monitoring system that enhances security and scalability management through a cloud-based federated learning environment. This system allows for effective monitoring and prediction of drug side effects while protecting the privacy of data shared between hospitals. The goal is to reduce mortality due to drug side effects and cut medical costs, exploring various technical approaches and methodologies to achieve this.

• The Journal of Bigdata
• /
• v.9 no.1
• /
• pp.101-113
• /
• 2024

Federated learning has garnered attention as an efficient method for training machine learning models in a distributed environment while maintaining data privacy and security. This study proposes a novel FedRFBagging algorithm to optimize the performance of random forest models in such federated learning environments. By dynamically adjusting the trees of local random forest models based on client-specific data characteristics, the proposed approach reduces communication costs and achieves high prediction accuracy even in environments with numerous clients. This method adapts to various data conditions, significantly enhancing model stability and training speed. While random forest models consist of multiple decision trees, transmitting all trees to the server in a federated learning environment results in exponentially increasing communication overhead, making their use impractical. Additionally, differences in data distribution among clients can lead to quality imbalances in the trees. To address this, the FedRFBagging algorithm selects only the highest-performing trees from each client for transmission to the server, which then reselects trees based on impurity values to construct the optimal global model. This reduces communication overhead and maintains high prediction performance across diverse data distributions. Although the global model reflects data from various clients, the data characteristics of each client may differ. To compensate for this, clients further train additional trees on the global model to perform local optimizations tailored to their data. This improves the overall model's prediction accuracy and adapts to changing data distributions. Our study demonstrates that the FedRFBagging algorithm effectively addresses the communication cost and performance issues associated with random forest models in federated learning environments, suggesting its applicability in such settings.