Earticle

본 연구는 다중회귀분석과 베이지안 네크워크를 이용해 수도권 상가 낙찰가율에 영향을 미치는 요인을 실증분석하였다. 종속변수는 상가낙찰가율로 독립변수는 상가낙찰율, 회사 채수익률, 산업생산지수, 상업용 건축물착공현황과 경제심리지수로 설정하였다, 시간적 범 위는 2003년 1월부터 2019년 12월까지로 하였고 공간적 범위는 수도권인 서울, 경기, 인천 으로 설정하였다. 실증분석결과, 서울시 상가낙찰가율에 상업용 건축물착공현황이 가장 많은 영향을 미치고 경기도 상가낙찰가율에는 회사채수익률이 가장 많은 영향을 미치고 경 제심리지수도 많은 영향을 미치는 것을 알 수가 있었다. 인천 상가낙찰가율에는 경기도 상 가낙찰가율이 영향을 가장 많이 미치고 산업생산지수가 그 다음으로 많은 영향을 미치는 것으로 나타났다. 지역별로 상가 낙찰가율에 영향을 미치는 요인은 차이가 있지만 전반적 으로 경제성장, 공급과 유동성이 많은 영향을 미치고 인근 지역의 낙찰가율도 영향을 미치는 것으로 알 수 있었다.

This study empirically analyzed the factors influencing the retail property aution price ratio in the metropolitan area using multiple regression analysis and Bayesian network approach. The dependent variable was set as the auction price ratio, and the independent variables were set as the auction ratio, corporate bond yield, industrial production index, commercial building construction, and economic sentiment index. The time range was from January 2003 to December 2019, and the spatial range was Seoul, Gyeonggi and Incheon. It was found that commercial building construction had the most influence in Seoul, corporate bond yield had the most influence in Gyeonggi-do, and the economic sentiment index also had a great influence. It was found that the Gyeonggi-do auction price ratio had the most influence on the Incheon auction price ratio, followed by the industrial production index. Although there are differences in factors affecting the auction price ratio for each region, it was found that overall, economic growth, supply and liquidity have a great influence, and the auction price ratio in the neighbor region also has an effect.

1인 창조기업에 대한 정부의 지원이 확대되고, 인공지능과 IT 기술의 발달로 1인 창조기업의 성장할 수 있는 환경이 좋아졌다. 이로 인해 1인 창조기업에 대한 사람들의 관심이 증가하고, 1인 창조기업 지원센터를 찾는 청년 창업가의 숫자도 크게 증가하고 있다. 본 연구는 비지도 기계학습과 베이지안 네트워크를 결합한 새로운 유형의 확률적 구조방정식 모형, 즉 PSEM을 사용하여 국내 1인 창조기업의 당기순이익에 영향을 주는 요인을 분석하고자 한다. 아울러 해당 영향요인들이 국내 1인 창조기업의 당기순이익에 미치는 인과관계 값을 산출하기 위하여 What-If 분석을 실시하였다. 분석 결과 당기 순이익의 변화에 큰 영향을 받지 않는 요인들은 서비스 업종일 때 ‘초기 자본금’, ‘창업자특성’, ‘준비 기간’, ‘창업자 배경’, ‘회사 특성’으로 나타났다. 서비스 업종을 제외하였을 때는 ‘창업자 배경’, ‘인식’, ‘초기 자본금’, ‘회사 특성’, ‘창업자특성’으로 분석되었다. 당기순이익에 영향을 주는 요인은 ‘수익’, ‘종업원’, ‘판매’임을 발견하였다.

In the era of economic recession caused by covid-19 pandemic, government expanded various types of supports for one-man creative companies with a bid to promoting healthier ecosystem for start-ups. Furthermore, the 4th industrial revolution atmospheres emboldened by AI technologies have countenanced young entrepreneurs’attempts to launch their own one-man creative companies. However, despite positive trends with one-man creative companies like this, few studies exist investigating which factors affect its net profit. To fill the research void, this study proposes a novel method of probabilistic structural equation model (PSEM) combining unsupervised machine learning and Bayesian networks. Empirical results revealed that those influential factors affecting one-man creative companies’ net profit encompass ‘income’, ‘number of employees’, and ‘sales’. On the contrary, multiple factors initially deemed to be significant to the net profit were excluded, like “background of the CEO”, “characteristics of the company”, “initial fund”, ‘preparation’, and “characteristics of the CEO” in service company. The others company’s results showed that ‘background of the CEO’, ‘perception’, ‘initial fund’, ‘characteristics of the company’, and ‘characteristics of the CEO’ were not included in factors affected by changes of net profit.

For the personalized learning, a good testing method, which can effectively estimate a learner’s proficiency, is required. In this paper, we propose a novel testing method, Bayesian network-based approach to Computerized Adaptive Testing (CAT). Our novel approach can estimate proficiency of the examinee effectively and efficiently because it reflects complicated relationships between all items and their categories, and can estimate detailed proficiency about each specific category. In experimental results, we show that our approach can improve accuracy and speed of estimating examinee’s proficiency as compared with classical testing methods like paper-based test and conventional IRT-based CAT.

Launch Vehicle software has played an important role in Launch Vehicle system. However, the reliability assessment of Launch Vehicle software is still a hard problem due to the complexity of Launch Vehicle software. Failure prediction can be an effective approach of Launch Vehicle reliability evaluation, whereas failure prediction of software has not yet been fully explored. In this paper, a Markov Bayesian Network model for Launch Vehicle software failure prediction (MBNG) is proposed. In MBNG, unique features of Launch Vehicle software is considered as an important part in the modeling process, which improves the effectiveness of this novel model. Experiments are conducted to verify the effectiveness of MBNG model and compare its performance with classic models.

Ubiquitous decision support systems have remained an imaginary and almost useless system for decades since its first introduction in early 1990’. However, it came out of lab into real world as ubiquitous computing became tangible in the form of mobile devices, pervasive mechanisms, and various mobile Internet technologies. Typically, context-aware systems had received acclaims from both researchers and practitioners as an alternative to making ubiquitous systems touch-and-feel electronics to the users. Nevertheless, context-aware systems lack predictive power which is essential for any ubiquitous systems to suggest timely and effective information for users. Poorly predicted information is likely to degrade the ubiquitous systems seriously. In this respect, context prediction mechanism emerges as a reliable vehicle for making ubiquitous systems more sustainable decision support tool for users. Despite the potentials of context prediction mechanism, few reliable mechanisms exist in literature which shows robust performance against changes in user’ contexts. For this reason, we propose a new type of ubiquitous decision support system that is powered by General Bayesian Network (GBN) capable of organizing causal relationships among a set of related variables. Drawing on the GBN’ strengths, this study proposes U-BASE (Ubiquitous Bayesian network-Assisted Support Engine) to suggest more reliable solution for the context prediction tasks. Performance of U-BASE was tested against real contextual data set, garnering very robust results. The practical implications are fully discussed with some future research issues.

This paper investigates a probabilistic causation problem that causes increase the probabilities of their effects on a suitable probabilistic model. A Bayesian network is used to analyze causal relationships of questions in Korean alcoholism questionnaire survey. We made a Korean obsessive compulsive drinking scale questionnaire for college students, and collected 73 responses to 21 questions over the course of an interview with investigating personnel from three different groups of responders. To identify and justify the casual relationships among questions we built an intial Bayesian network with an expert point of view’s network structure. However, G-test analysis of the variables, survey questions, with sample data revealed several discrepancies from the initial Bayesian network regarding relationship between nodes and also introduced an isolated subgraph. Interestingly, 3 potential arcs were identified and used to construct a modified Bayesian network which best correspond to the intention of the alcoholism questionnaire. Future works will assess the significance of these potential arcs with larger data sets and investigate whether revision of the alcoholism questionnaire is necessary for better diagnosis.

A novel idea is presented to locate intersections in a video sequence captured from a moving vehicle. More specifically, we propose a Bayesian network approach to combine evidence extracted from a video sequence and evidence from a database, maximizing evidence from various sensors in a systematic manner and locating intersections robustly.

OBJECTIVES: Recent studies have shown that the rate of needlestick and sharps injuries (NSIs) is unacceptably high in Iranian hospitals. The aim of the present study was to use a systematic approach to predict and reduce these injuries. METHODS: This cross-sectional study was conducted in 5 hospitals in Tehran, Iran. Eleven variables thought to affect NSIs were categorized based on the Human Factors Analysis and Classification System (HFACS) framework and modeled using a Bayesian network. A self-administered validated questionnaire was used to collect the required data. In total, 343 cases were used to train the model and 50 cases were used to test the model. Model performance was assessed using various indices. Finally, using predictive reasoning, several intervention strategies for reducing NSIs were recommended. RESULTS: The Bayesian network HFACS model was able to predict 86% of new cases correctly. The analyses showed that safety motivation and fatigue were the most important contributors to NSIs. Supervisors’ attitude toward safety and working hours per week were the most important factors in the unsafe supervision category. Management commitment and staffing were the most important organizational-level factors affecting NSIs. Finally, promising intervention strategies for reducing NSIs were identified and discussed. CONCLUSIONS: To reduce NSIs, both management commitment and sufficient staffing are necessary. Supervisors should encourage nurses to engage in safe behavior. Excessive working hours result in fatigue and increase the risk of NSIs.

Since online game industry keeps growing fast, online game players’ behaviors have become interesting topic for researchers. Prior researches mostly measured attitude, flow, satisfaction, perceived enjoyment and playfulness to describe online games players’ behaviors. Our study investigates psychologically negative factors like perceived loneliness and stress which haven’t been attracted much attention. With positive factors such as character identification and enjoyment, this study revealed perceived stress significantly affects flow which directly affect online game players’ loyalty. In addition, we shed light on methodological approach of Bayesian Network(BN) structuring to describe the relationship between factors. After empirical approach with Structural Equation Model(SEM) to test our hypotheses, we apply various types of BN. Combining partially subsets of BN(CBN) with each other, a better performing whole structure of nodes could be built. Moreover, we conducted structuring a BN based on SEM(SBN). SBN is found to have some advantages of predicting performance and help managerial decision by what-if diagnosis. Other implications of our approaches, limitation, and suggestions for further researches are also discussed at the end of this study.

본 연구는 전통적인 역사학 연구에 컴퓨테이셔널 방법론을 접목한 디지털 역사학적 접근을 통해 고대 한국 신라 지배층의 권력 구조를 분석한다. 질적 연구 방법은 역사적 의미와 맥락을 해석하는 데 유효하지만, 고대 사회의 역동적인 관계 구조를 복원하는 데에는 한계가 있다. 특히 역사 기록은 종종 단편적이고 불완전하다. 따라서 변수 간 조건부 의존성을 모델링할 수 있는 베이지안 네트워크는 불확실하고 부분적으로 누락된 데이터를 해석하는 데 효과적인 분석 도구이다. 본 연구에서는 사료에 등장하는 주요 인물들의 친족 관계, 관직 소속, 연대 정보를 체계화하여 구조화된 데이터셋을 구축하였다. 이를 바탕으로 베이지안 네트워크 모델을 개발하고 확률적 추론을 적용하여 누락된 가족 관계를 식별하였다. AI 기반 네트워크 분석을 활용해 신라 지배층의 친족 관계망을 재구성하였다. 분석 결과, 골품제, 동일 성씨, 동일 이름자가 출신, 즉 본인의 가족과 배우자의 가족을 추론하는 데 핵심적인 역할을 하는 것으로 나타났다. AI 기반 베이지안 네트워크 모델의 구축과 분석을 통해 신라 지배층의 계보 구조와 권력 관계를 부분적으로 복원하였다. 베이지안 네트워크 모델의 가장 큰 의의는 특정 시기나 주제에 국한되지 않고 보다 보편적인 적용 가능성을 지닌다는 점에 있다. 본 연구는 지배층의 친족 네트워크를 기반으로 한 권력 분포 양상을 이해하는 데 있어 AI 기반 베이지안 네트워크 분석이 유효한 연구 방법론임을 입증하며, 나아가 디지털 역사학이라는 새로운 연구 방법의 가능성을 제시함으로써 역사학 연구의 지평을 확장한다.

This study analyzes the power structure of the ruling elite in ancient Korea’s Silla kingdom through a digital history approach that integrates computational methodologies into traditional historical research. While qualitative methods are effective for interpreting historical meanings and contexts, they have limitations in reconstructing the dynamic relational structures of ancient societies. In particular, historical records are often fragmentary and incomplete. Therefore, Bayesian networks, which can model conditional dependencies among variables, serve as an effective analytical tool for interpreting uncertain and partially missing data. In this study, a structured dataset was constructed by systematizing kinship relations, bureaucratic affiliations, and chronological information of key historical figures appearing in the sources. Based on this, a Bayesian network model was developed and probabilistic inference was applied to identify missing familial relationships. Using AI-based network analysis, the kinship networks of the Silla ruling elite were reconstructed. The analysis revealed that elements such as the bone-rank system, shared surnames, and common generational name characters played a crucial role in inferring both the natal family and the spouse’s family. Through the construction and analysis of an AI-based Bayesian network model, this study partially reconstructed the genealogical structure and power relations of the ruling elite in Silla. The most significant value of the Bayesian network model lies in its applicability beyond a specific period or subject. This research demonstrates that AI-based Bayesian network analysis is an effective methodology for understanding patterns of power distribution grounded in kinship networks among the ruling class. Furthermore, it presents the potential of digital historical research as a new methodological horizon, thereby expanding the scope of historical studies

Recent advances in artificial intelligence (AI) have rapidly transformed the landscape of modern warfare, with AI-enabled weapon systems offering unprecedented improvements in target recognition, operational speed, and autonomous mission execution. However, despite these benefits, the reliability of autonomous decision-making remains a major concern, particularly in certain environments where incorrect predictions can result in unintended engagements or civilian casualties. Addressing this issue requires a means to not only improve prediction accuracy but quantify the uncertainty behind each decision. This study proposes a novel methodology for evaluating and enhancing the reliability of autonomous decisions in AI weapon systems by leveraging the framework of Bayesian Neural Networks(BNNs). Recognizing the computational limitations of applying BNNs to an entire deep learning model, a Partially Bayesian Neural Network (PBNN) architecture is introduced. This design enables the system to estimate the uncertainty of its own outputs while maintaining real-time performance-an essential requirement in military applications. Furthermore, a semi-autonomous control structure is implemented, in which human intervention is selectively triggered based on the estimated uncertainty. The results show that the proposed method produces significantly lower output variance and higher confidence scores when predictions are correct, while consistently identifying high-uncertainty cases associated with incorrect predictions. This allows the system to distinguish between reliable and unreliable decisions in real time. The proposed approach offers a technically viable pathway toward balancing autonomy and human oversight in AI weapon systems. It also serves as a foundational framework for the future implementation of Manned-Unmanned Teaming (MUM-T).