Bong-Gyu Jang 보험학회지

[논문] 국가유산보험의 이득금지 원칙 위반 가능성에 대한 고찰

Abstract 우리 정부는 국가유산기본법에 의거하여 국가유산에 발생하는 손해를 전부 무상으로 수리하여 주는 국가유산수리제도를 운영하고 있다. 따라서 국가유산의 소유자가 국가유산을 목적물로 하는 보험에 가입하면 보험자의 보험금 지급 거부 및 대위권 행사 등으로 인한 법적 분쟁이 발생할 가능성이 있다. 또한 국가유산 소유자가 보험금과 국가유산수리제도로부터 발생한 손해를 초과하는 보상(초과보상)을 받게 될 경우 이득금지 원칙이 위배되고, 그렇지 않은 경우 보험의 효용이 없어지는 딜레마가 발생한다. 이와 같은 문제를 해소하기 위하여 국가유산보험과 국가유산수리제도의 관계 정립 및 국가유산수리제도의 정비가 필요하다.

Kwangmin Jung North American Actuarial Journal

[논문] Spatial cyber loss clusters at county level and socio-economic determinants of cyber risks

Abstract This study investigates whether cyber loss events occurring in the United States are spatially correlated and if so, which socioeconomic factors are associated with the spatial correlation. We analyze 3132 counties of the 50 U.S. states from 2005 to 2020 using the largest existing dataset of cyber risks and socioeconomic data. While previous literature found no or little spatial correlation at the state level, we are the first to document that such correlation exists at the county level; positive Moran’s I indicates that more exposed (i.e., a relatively large number of cyber events and losses) and less exposed counties are clustered. Spatial regressions show positive direct and negative indirect effects of county-level population and average income on loss frequency and severity. Large and wealthy counties thus tend to be more exposed to cyber risk events, but their geographically neighboring counties are less affected. We further investigate relatively exposed regions (California and the Northeast Coast) and three risk types (malicious, unintended, and privacy risks) and show consistent spatial effects for the key variables of population size and average income. Our findings can aid risk managers, cyber insurers, and policymakers to geographically differentiate cyber risk, recognize relatively more exposed regions, and develop more effective risk management strategies.

Kwang Jae Kim IEEE Transactions on Reliability

[논문] Set Response Surface Methodology and its Application in Solving the Wrinkle and Crack Problem in the Auto Industry

Abstract This study considers a response surface methodology (RSM) variation in which a response has multiple central tendencies (MCTs) that can have multiple influences on a sheet metal part. This is formulated as a set response surface (SRS) problem, which, in industrial practice, is studied using the thinning ratio example. The set RSM (SRSM), consisting of three phases, is proposed to solve the SRS problem.The first phase is the problem definition and regional division phase, where based on the analysis of MCTs and response influence, a sheet metal part that needs quality improvement is divided into q regions. The second phase is the experiment and data regression phase. The third phase is the optimization and interactive decision phase, where the condition relaxation strategy (CRS) is proposed, and relaxation is obtained based on the barrier and engineering requirement analysis.The optimization models are constructed for the (q+1)-level optimal solutions using the CRS. The proposed SRSM is verified by tests on the wrinkle and crack problem of the inner plate of the back door. Apossible optimization model combination and trend analysis strategy is proposed to solve the CRS challenge for higher dimensions in the tendencies.

Sung H. Han International Journal of Human-Computer Interaction

[논문] Emotional Experience of Audiences in 4D Content

Abstract This study aims to identify and classify the factors influencing the emotionalexperience of audiences engaging with four-dimensional (4D) content. Throughsystematic literature reviews, expert workshops, and experiments, factors wereidentified and refined. Factors were first identified through a systematic literaturereview, then expanded with expert insights on 4D content design and userexperience. Finally, audience experiences were analyzed through interviews followingexperiments with 32 participants, further refining the factors. The 151 collectedfactors were categorized based on common characteristics, resulting in the identification of 5 major groups and 25 subgroups. This classification systemhighlights the diversity of factors such as content experience, viewing environment,sensory stimulus design, personal characteristics, and content composition,emphasizing the importance of combining these elements to enhance emotionalengagement. Furthermore, it provides a valuable theoretical foundation and practicalguidance for 4D content creators, enabling them to design more immersive andemotionally resonant experiences.

Minwoo Chae Journal of the Korean Statistical Society

[논문] Rates of convergence for nonparametric estimation of singular distributions using generative adversarial networks

Abstract It is common in nonparametric estimation problems to impose a certain low-dimensional structure on the unknown parameter to avoid the curse of dimensionality. This paper considers a nonparametric distribution estimation problem with a structural assumption under which the target distribution is allowed to be singular with respect to the Lebesgue measure. In particular, we investigate the use of generative adversarial networks (GANs) for estimating the unknown distribution and obtain a convergence rate with respect to the L1-Wasserstein metric. The convergence rate depends only on the underlying structure and noise level. More interestingly, under the same structural assumption, the convergence rate of GAN is strictly faster than the known rate of VAE in the literature. We also obtain a lower bound for the minimax optimal rate, which is conjectured to be sharp at least in some special cases. Although our upper and lower bounds for the minimax optimal rate do not match, the difference is not significant.

Minseok Song/Minwoo Chae Journal of the Korean Statistical Society

[논문] Group-constrained latent Dirichlet allocation for fashion data analysis

Abstract Recent advances in machine learning have provided valuable tools for constructing various recommendation systems in e-commerce companies such as Amazon and eBay. In this paper, we analyze click history records from an online fashion mall using a well-known Bayesian topic model, the latent Dirichlet allocation (LDA). Although multinomial mixture models such as the LDA have been widely applied to analyze such count data, a basic LDA-based approach for fashion data analysis may yield a crucial issue. For a customer who clicked pants primarily, for example, a basic recommendation algorithm tends to recommend pants only. Given a click history of pants, a more desirable algorithm would recommend fashion items compatible with the clicked pants. For this purpose, we propose a novel Bayesian model called the group-constrained LDA, which can incorporate prior information about the item groups. The proposed method is applied to analyze the click history data from Samsung Fashion (SSF) Shop, one of the largest online fashion malls in South Korea.

Young Myoung Ko OR Spectrum

[논문] Queueing management for reducing the overlaps of customers in service systems

Abstract Implementing effective queueing management strategies to reduce the gathering level of customers has profound implications in reducing customers’ infection risk during pandemic seasons and improving the service experience. This paper investigates how different queueing topologies and flow control policies impact customers’ gathering within service systems from a queueing modeling perspective. To model the gathering levels, we consider two metrics: average overlapping time and average number of overlapped customers. We prove that both metrics possess a symmetry property, enabling us to derive closed-form expressions for accurate evaluation. We focus on two commonly used queueing topologies: serial topology in which customers travel through the queueing zones sequentially, and the parallel topology in which customers receive service in separated queues. For serial topology systems, we propose practical flow control schemes that are easily applicable to service systems. For the parallel topology system, we find that while even splitting creates independent service zones, it fails to achieve shorter overlapping times compared to the serial topology system due to the loss in service efficiency. However, by employing the join-the-shortest-queue scheme, the parallel system can better utilize the waiting zones, resulting in smaller overlapping times and numbers compared to the serial topology system. This emphasizes the importance of routing in the parallel system for reducing customers’ overlaps. Furthermore, we discuss the grocery model, the impact of service distributions on policy design, and the overlaps between customers and staff, to provide a more comprehensive understanding of how to design effective management strategies when considering overlaps.

Bong-Gyu Jang Applied Economics

[논문] Forecasting realized volatility of the oil future prices via machine learning

ABSTRACT This paper explores the potential use of machine learning models in crude oil realizedvolatility forecasting through a variety of empirical analyses and robustness checks.Although the conventional Heterogeneous Autoregressive (HAR) model is widelyaccepted, the machine learning models with the HAR factors can significantly improveits forecasting performance. We also found that macroeconomic variables such assupply factors, implied volatility indices and uncertainty factors can be useful inforecasting oil volatility.