머신러닝 기반 기업부도위험 예측모델 검증 및 정책적 제언: 스태킹 앙상블 모델을 통한 개선을 중심으로

엄하늘(Eom, Haneul),김재성(Kim, Jaeseong),최상옥(Choi, Sangok) 한국지능정보시스템학회 2020 지능정보연구 Vol.26 No.2

This study uses corporate data from 2012 to 2018 when K-IFRS was applied in earnest to predict default risks. The data used in the analysis totaled 10,545 rows, consisting of 160 columns including 38 in the statement of financial position, 26 in the statement of comprehensive income, 11 in the statement of cash flows, and 76 in the index of financial ratios. Unlike most previous prior studies used the default event as the basis for learning about default risk, this study calculated default risk using the market capitalization and stock price volatility of each company based on the Merton model. Through this, it was able to solve the problem of data imbalance due to the scarcity of default events, which had been pointed out as the limitation of the existing methodology, and the problem of reflecting the difference in default risk that exists within ordinary companies. Because learning was conducted only by using corporate information available to unlisted companies, default risks of unlisted companies without stock price information can be appropriately derived. Through this, it can provide stable default risk assessment services to unlisted companies that are difficult to determine proper default risk with traditional credit rating models such as small and medium-sized companies and startups. Although there has been an active study of predicting corporate default risks using machine learning recently, model bias issues exist because most studies are making predictions based on a single model. Stable and reliable valuation methodology is required for the calculation of default risk, given that the entitys default risk information is very widely utilized in the market and the sensitivity to the difference in default risk is high. Also, Strict standards are also required for methods of calculation. The credit rating method stipulated by the Financial Services Commission in the Financial Investment Regulations calls for the preparation of evaluation methods, including verification of the adequacy of evaluation methods, in consideration of past statistical data and experiences on credit ratings and changes in future market conditions. This study allowed the reduction of individual models bias by utilizing stacking ensemble techniques that synthesize various machine learning models. This allows us to capture complex nonlinear relationships between default risk and various corporate information and maximize the advantages of machine learning-based default risk prediction models that take less time to calculate. To calculate forecasts by sub model to be used as input data for the Stacking Ensemble model, training data were divided into seven pieces, and sub-models were trained in a divided set to produce forecasts. To compare the predictive power of the Stacking Ensemble model, Random Forest, MLP, and CNN models were trained with full training data, then the predictive power of each model was verified on the test set. The analysis showed that the Stacking Ensemble model exceeded the predictive power of the Random Forest model, which had the best performance on a single model. Next, to check for statistically significant differences between the Stacking Ensemble model and the forecasts for each individual model, the Pair between the Stacking Ensemble model and each individual model was constructed. Because the results of the Shapiro-wilk normality test also showed that all Pair did not follow normality, Using the nonparametric method wilcoxon rank sum test, we checked whether the two model forecasts that make up the Pair showed statistically significant differences. The analysis showed that the forecasts of the Staging Ensemble model showed statistically significant differences from those of the MLP model and CNN model. In addition, this study can provide a methodology that allows existing credit rating agencies to apply machine learning-based bankruptcy risk prediction methodologies, given that traditional credit rating models can a

자연과학 분야 연구 데이터 리포지터리 속성분석 : 운영수준, 주제, 유형, 활용도를 중심으로

김우중(Woojoong Kim),김형준(Hyungjoon Kim),박희진(Heejin Park),최상옥(Sangok Choi) 한국기술혁신학회 2020 한국기술혁신학회 학술대회 발표논문집 Vol.2020 No.11

본 연구는 향후 데이터 관리계획을 수립하여야 할 출연연 등 관계기관에 시사점을 주기 위하여 오픈 액세스와 혁신확산이론, 사회연결망 이론을 바탕으로 Re3data.org 내 402개의 자연과학 분야 연구 데이터 리포지터리의 세부 속성을 대상으로 기술통계분석, 계층적 군집분석, 대응일치분석, 네트워크 분석을 시행하였다. 분석결과, 연구 데이터의 파급력을 높이기 위해서는 방문 빈도가 높은 연구 데이터 리포지터리의 세부 속성을 추가 분석할 필요가 있고, 외부 데이터의 활용도 제고를 위해서는 평가제도 개선, 인식 개선 등을 위한 폭넓은 정책연구가 필요할 것으로 보인다. Based on the Open Access Theory, Innovation Diffusion Theory and Social Network Theory, this study conducted descriptive statistics, hierarchical clustering analysis, correspondence analysis and network analysis on the different variables from 402 Natural Sciences’ research data repositories in Re3data.org in order to give policy implications to related institutes and agencies that need to establish data management plans in the future. In order to increase influence of one’s data, it seems necessary to further analyze more variables of the frequently visited research data repositories, and in order to enhance the utilization of others’ data, a wide range of policy studies such as improving the evaluation system of researchers or raising awareness are needed.

자기회귀교차지연모형을 적용한 전략적 인적자원관리와 조직몰입 간의 종단적 관계 : 평가시스템 공정성의 매개효과 중심으로

김재성(Jaeseong Kim),이나래(Narae Lee),엄하늘(Haneul Eom),최상옥(Sangok Choi) 한국인사관리학회 2020 조직과 인사관리연구 Vol.44 No.2

지금까지 횡단적으로 분석되었던 전략적 인적자원관리와 종업원의 조직몰입의 관계는 시차를 고려한 인과관계를 바탕으로 종단적으로 해석해야 할 필요성이 있다. 이를 위해 인적자본기업패널(HCCP) 종단자료를 사용하여 전략적 인적자원관리 관점에서 인사부문의 전략적 참여 확대와 인사제도 중 하나인 평가시스템의 공정성이 조직몰입이라는 성과에 시간의 흐름에 따라 영향을 미치는지 자기회귀교차지연모형(Autoregressive Cross-Lagged Modeling)을 활용하여 실증 분석하였다. 연구결과, 인사부문의 전략적 참여와 평가시스템의 공정성은 상호 간에 긍정적인 영향을 주는 것으로 확인되었으며, 평가시스템의 공정성과 조직몰입 간에도 상호 긍정적인 영향이 확인되었다. 그리고 조직몰입이 인사부문의 전략적 참여에 영향을 주는 순환적 관계가 확인되었고, 평가시스템의 공정성이 인사부문의 전략적 참여와 조직몰입 간에 미치는 매개효과가 확인되었다. 본 연구는 전략적 인적자원관리 하에서 평가시스템의 공정성 변수가 전략과 조직몰입 사이에 영향을 미친다는 점을 제시하였고, 기업의 경영전략이 반영된 인사전략이 조직성과로 이어지기 위해서는 공정성이 담보된 인사제도의 구축과 시행이 중요하다는 것을 보여주고 있다. This study is an empirical study using the Autogressive Cross-Lagged Modeling on whether the HR involvement(HI) and the Procedural Justice of the Performance Evaluation System(PJ) in Strategic Human Resource Management(SHRM), affects the Organizational Commitment(OC). The result shows that between HI and PJ, PJ and OC have positive effects on each other. In addition, the cyclical causal relationship in which OC affects HI was identified, and the mediating effect of PJ on HI and OC was verified. This study demonstrated that the justice variables of the performance evaluation system in SHRM affect between strategy establishment and the organizational commitment. Therefore, the management has to consider the justice-guaranteed HR system to ensure that the HR strategy reflected by the management strategy of the company leads to organizational performance.

자연과학 분야 연구 데이터 리포지터리 속성분석 : 운영수준, 주제, 유형, 활용도를 중심으로

김우중(Woojoong Kim),김형준(Hyungjoon Kim),박희진(Heejin Park),최상옥(Sangok Choi) 한국기술혁신학회 2021 기술혁신학회지 Vol.24 No.4

본 연구는 과학기술 분야의 개방과 공유를 통한 개방형 혁신을 추구해야 할 정부출연연구기관 등 관계기관에 시사점을 주기 위하여 오픈 액세스와 사회연결망 이론을 바탕으로 Re3data.org 내 402개의 자연과학 분야 연구 데이터 리포지터리의 세부 속성을 대상으로 기술통계분석, 계층적 군집분석, 대응일치분석, 네트워크 분석을 시행하였다. 분석결과, 데이터 리포지터리의 운영수준에 대하여 ‘우수’, ‘보편’, ‘열악’, ‘비공개’의 4가지 유형의 클러스터를 도출하였으며, 연구 데이터 리포지터리의 연구주제에 따라 컨텐츠 타입에 차이가 있음을 확인하였다. 또한 북미와 유럽 등 특정 지역에서 데이터 리포지터리를 활발하게 이용하고 있음을 알 수 있었으며, 방문 빈도가 높은 데이터 리포지터리일수록 표준화된 연구문서 형태로 데이터가 공유되는 경우가 많다는 사실을 발견하였다. 개방형 혁신을 위해서는 데이터 리포지터리의 활용률을 제고할 필요가 있다. 이를 위하여 연구분야별 차별화된 데이터 활용 정책이 필요하고, 데이터 공유와 관련된 인식개선 및 인식이 실천으로 이어질 수 있는 동기부여 방안 마련이 필요해 보인다. Based on the Open Access Theory and Social Network Theory, this study conducted descriptive statistics, hierarchical clustering analysis, correspondence analysis and network analysis on the different variables from 402 Natural Sciences’ research data repositories in Re3data.org in order to give policy implications to government-funded research institutes and related agencies that need to pursue open innovation through openness and sharing of science and technology. As a result of the analysis, four types of clusters were derived for the level of operation of the data repository and the content type was found dependent on the research topic of the research data repository. We also found that data repositories are actively used in certain regions, such as North America and Europe, and that the more frequently visited data repositories are often shared in the form of standardized research documents. To increase utilization of data repositories, differentiated data utilization policies are needed in each research field, and raising awareness of data sharing’s importance, including motivational measures that can lead to action, is needed.

초산 생산을 위한 폐기물 가스화 합성가스 생산 특성

구재회(Jaehoi Gu),박수남(Soonam Park),김나랑(Narang Kim),임용택(Yongtaek Lim),박영수(Yeongsu Park),윤영식(Youngsik Yoon),남성방(Sung-Bang Nam),전병준(Byungjun Chon),박혜정(Haejoung Park),최상옥(Sangok Choi),추수태(Sootae Choo) 한국신재생에너지학회 2014 한국신재생에너지학회 학술대회논문집 Vol.2014 No.05