• Journal of Intelligence and Information Systems
• /
• v.26 no.2
• /
• pp.105-129
• /
• 2020

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 entity's 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 also be reflected as sub-models to calculate the final default probability. Also, the Stacking Ensemble techniques proposed in this study can help design to meet the requirements of the Financial Investment Business Regulations through the combination of various sub-models. We hope that this research will be used as a resource to increase practical use by overcoming and improving the limitations of existing machine learning-based models.

• Journal of Korean Society of Forest Science
• /
• v.90 no.2
• /
• pp.197-209
• /
• 2001

TOPMODEL, semi-distributed hydrological model, is frequently applied to predict the amount of discharge, main flow pathways and water quality in a forested catchment, especially in a spatial dimension. TOPMODEL is a kind of conceptual model, not physical one. The main concept of TOPMODEL is constituted by the topographic index and soil transmissivity. Two components can be used for predicting the surface and subsurface contributing area. This study is conducted for the validation of applicability of TOPMODEL at small forested catchments in Korea. The experimental area is located at Gwangneung forest operated by Korea Forest Research Institute, Gyeonggi-do near Seoul metropolitan. Two study catchments in this area have been working since 1979 ; one is the natural mature deciduous forest(22.0 ha) about 80 years old and the other is the planted young coniferous forest(13.6 ha) about 22 years old. The data collected during the two events in July 1995 and June 2000 at the mature deciduous forest and the three events in July 1995 and 1999, August 2000 at the young coniferous forest were used as the observed data set, respectively. The topographic index was calculated using $10m{\times}10m$ resolution raster digital elevation map(DEM). The distribution of the topographic index ranged from 2.6 to 11.1 at the deciduous and 2.7 to 16.0 at the coniferous catchment. The result of the optimization using the forecasting efficiency as the objective function showed that the model parameter, m and the mean catchment value of surface saturated transmissivity, $lnT_0$ had a high sensitivity. The values of the optimized parameters for m and InT_0 were 0.034 and 0.038; 8.672 and 9.475 at the deciduous and 0.031, 0.032 and 0.033; 5.969, 7.129 and 7.575 at the coniferous catchment, respectively. The forecasting efficiencies resulted from the simulation using the optimized parameter were comparatively high ; 0.958 and 0.909 at the deciduous and 0.825, 0.922 and 0.961 at the coniferous catchment. The observed and simulated hyeto-hydrograph shoed that the time of lag to peak coincided well. Though the total runoff and peakflow of some events showed a discrepancy between the observed and simulated output, TOPMODEL could overall predict a hydrologic output at the estimation error less than 10 %. Therefore, TOPMODEL is useful tool for the prediction of runoff at an ungaged forested catchment in Korea.

• Asia-Pacific Journal of Business Venturing and Entrepreneurship
• /
• v.13 no.2
• /
• pp.63-77
• /
• 2018

In the rapidly to change global market in recent years, as the era of merging and integrating industries and the evolution of technology have come to an era in which everything can not be solved as a single company, it is evolving into competition for the enterprise network rather than the competition for the enterprise unit. In a competitive business environment, it is necessary to provide not only for the efforts as an individual companies but also the mutual development efforts to enhance output through the innovation activities based on the interrelationship with the business partners. In spite of the recent efforts and research through core competencies and innovation activities, some of business activities were unable to achieve enough progress in business performance and this study mainly focused to improve business performance for those companies. This study targeted CEOs and Directors who participates in "manufacturing performance innovation partnership project" carried by The foundation of Large, SMEs, Agriculture, Fisheries cooperation Korea and studied the influences of innovation activities to the core competencies and business performance. Detailed variables in this study were extracted from the previous research and used for verification. The study is designed to determine the influence of individual innovation activities to the core competencies and business performance. Innovation activities as a parameter, the relationship between core competencies and business performance was examined. In the examination of the innovation activities as a meditated effect, those activities carried by SMEs (Collaboration in Technology, Manufacturing, and Management innovations with Large Scale Business) through partnership in manufacturing innovation is significantly related business performance. Therefore, the result reveals that the individual SMEs are having own limitation in the achievement of significant progress in business performance with their own capabilities, and using the innovation activities act as catalyst through the collaboration with large scale businesses would result significant progress in business performance. Mutual effort in collaborative innovation activities between large scale businesses and SMEs is one of the most critical issues in recent years in Korea and the main focus of this study is to provide analysis which demonstrates where the SMEs are required to focus in their innovation activities.

• Progress in Medical Physics
• /
• v.20 no.4
• /
• pp.235-243
• /
• 2009

The purpose of this experiment is to know the relation of the detection and characterization of liver's diseases as comparison of finding at MR imaging using a Ferucarbotran (SPIO) and Gd-EOB-DTPA (Primovist) agents in diffuse liver disease. A total of 50 patients (25 men and 25 women, mean age: 50 years) with liver diseases were investigated at 3.0T machine (GE, General Electric Medical System, Excite HD) "with 8 Ch body coil for comparison of diseases and contrast's uptake relation, which used the LAVA, MGRE." All images were performed on the same location with before and after Ferucarbotran and Gd-EOB-DTPA administrations (p<0.05). Contrast to noise ratio of Ferucarbotran and Gd-EOB-DTPA in the HCC were $3.08{\pm}0.12$ and $7.00{\pm}0.27$ with MGRE and LAVA pulse sequence, $3.62{\pm}0.13$ and $2.60{\pm}0.23$ in the hyper-plastic nodule, $1.70{\pm}0.09$ and $2.60{\pm}0.23$ in the meta, $2.12{\pm}0.28$ and $5.86{\pm}0.28$ in the FNH, $4.45{\pm}0.28$ and $1.73{\pm}0.02$ in the abscess and ANOVA test was used to evaluate the diagnostic performance of each disease (p<0.05). In conclusions, two techniques were well demonstrated with the relation of the detection and characterization of liver's diseases.

• Progress in Medical Physics
• /
• v.20 no.4
• /
• pp.290-297
• /
• 2009

In case of radiation treatment using small field high-energy photon beams, an accurate dosimetry is a challenging task because of dosimetrically unfavorable phenomena such as dramatic changes of the dose at the field boundaries, dis-equilibrium of the electrons, and non-uniformity between the detector and the phantom materials. In this study, the absorbed dose in the phantom was measured by using an ion chamber and a diode detector widely used in clinics. $GAFCHROMIC^{(R)}$ EBT films composed of water equivalent materials was also evaluated as a small field detector and compared with ionchamber and diode detectors. The output factors at 10 cm depth of a solid phantom located 100 cm from the 6 MV linear accelerator (Varian, 6 EX) source were measured for 6 field sizes ($5{\times}5\;cm^2$, $2{\times}2\;cm^2$, $1.5{\times}1.5\;cm^2$, $1{\times}1\;cm^2$, $0.7{\times}0.7\;cm^2$ and $0.5{\times}0.5\;cm^2$). As a result, from $5{\times}5\;cm^2$ to $1.5{\times}1.5\;cm^2$ field sizes, absorbed doses from three detectors were accurately identified within 1%. Wheres, the ion chamber underestimated dose compared to other detectors in the field sizes less than $1{\times}1\;cm^2$. In order to correct the observed underestimation, a convolution method was employed to eliminate the volume averaging effect of an ion chamber. Finally, in $1{\times}1\;cm^2$ field the absorbed dose with a diode detector was about 3% higher than that with the EBT film while the dose with the ion chamber after volume correction was 1% lower. For $0.5{\times}0.5\;cm^2$ field, the dose with the diode detector was 1% larger than that with the EBT film while dose with volume corrected ionization chamber was 7% lower. In conclusion, the possibility of $GAFCHROMIC^{(R)}$ EBT film as an small field dosimeter was tested and further investigation will be proceed using Monte Calro simulation.

• Journal of Food Hygiene and Safety
• /
• v.29 no.3
• /
• pp.234-240
• /
• 2014

Fluxapyroxad is classified as carboxamide fungicide that inhibits succinate dehydrogenase in complex II of mitochondrial respiratory chain, which results in inhibition of mycelial growth within the fungus target species. This study was carried out to assure the safety of fluxapyroxad residues in agricultural products by developing an official analytical method. A new, reliable analytical method was developed and validated using High Performance liquid Chromatograph-UV/visible detector (HPLC-UVD) for the determination of fluxapyroxad residues. The fluxapyroxad residues in samples were extracted with acetonitrile, partitioned with dichloromethane, and then purified with silica solid phase extraction (SPE) cartridge. Correlation coefficient($R^2$) of fluxapyroxad standard solution was 0.9999. The method was validated using apple, pear, peanut, pepper, hulled rice, potato, and soybean spiked with fluxapyroxad at 0.05 and 0.5 mg/kg. Average recoveries were 80.6~114.0% with relative standard deviation less than 10%, and limit of detection (LOD) and limit of quantification (LOQ) were 0.01 and 0.05 mg/kg, respectively. All validation parameters were followed with Codex guideline (CAC/GL 40). LC-MS (Liquid Chromatograph-Mass Spectrometer) was also applied to confirm the analytical method. Base on these results, this method was found to be appropriate fluxapyroxad residue determination and can be used as the official method of analysis.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2006.05a
• /
• pp.720-725
• /
• 2006

토양수분은 식물의 생장 및 가용수자원 산정 등에 있어서 중요한 요소로서 토양층 상부의 수 m내에 존재하는 수분의 양을 일컫는다. 토양수분과 토양수분의 공간적 시간적 특징들은 증발, 침투, 지하수 재충전, 토양침식, 식생 분포 등을 지배하는 중요한 요소이다. 강우 등으로 인한 지면과 지표하에서의 순간적인 포화공간의 형성 및 유출의 생성 등을 포함하는 과정과 증발산 등은 모두 비포화대(vadose zone) 혹은 토양층에서의 토양수분의 함량에 크게 의존하게 된다(이가영 등(2005)). 분포형 수문모형은 유역을 격자단위로 세분화하여 매개변수를 부여하고, 증발산, 침투, 지표면유출, 중간유출, 지하수유출, 하도 흐름 등 여러 가지 수문요소를 해석하는 종합적인 수문모형이다. 지표면에 내린 강우가 증발, 침투, 유출될 지는 토양수분의 함량에 크게 의존하게 되며, 따라서 토양수분에 대한 적절한 모의가 분포형 수문모형의 정확도를 좌우하는 핵심이라 할 수 있다. 본 연구에서는 분포형 수문모형인 WEP 모형을 경안천 유역(유역면적: $575km^2$, 유로연장: 49.3㎞)에 적용하여 토양수분의 시공간분포를 모의하였다. 지점별 토양수분 모의결과, 토양 매개변수의 최대, 최소값 내에서 적절히 모의됨을 확인하였으나, 관측값이 없어 실질적으로 타당한지 여부는 검증하지 못하였다. 토양수분비율, 연간 증발산량, 지표면 유출량 공간분포를 비교한 결과, 토양수분비율이 연간 증발산량 모의에 직접적인 영향을 주는 것을 확인할 수 있었다. 일부격자에서는 토양수분이 지나치게 높게 모의되었는데, 지하수위와 관련있는 것으로 보이며, 구축된 자료가 부족한 지하대수층에 대한 정보부족이 토양수분 계산에도 영향을 준 것으로 보인다. 본 연구는 WEP 모형의 토양수분 해석능력에 대한 시험적용에 그 의의가 있으며, 향후 토양 및 지표하 매개변수 정보가 충분히 갖추어지고, 토양수분 관측결과 있는 대상유역에 대한 적용이 요구된다.-Moment 방법에 의해 추정된 매개변수를 사용한 Power 분포를 적용하였으며 이들 분포의 적합도를 PPCC Test를 사용하여 평가해봄으로써 낙동강 유역에서의 저수시의 유출량 추정에 대한 Power 분포의 적용성을 판단해 보았다. 뿐만 아니라 이와 관련된 수문요소기술을 확보할 수 있을 것이다.역의 물순환 과정을 보다 명확히 규명하고자 노력하였다.으로 추정되었다.면으로의 월류량을 산정하고 유입된 지표유량에 대해서 배수시스템에서의 흐름해석을 수행하였다. 그리고, 침수해석을 위해서는 2차원 침수해석을 위한 DEM기반 침수해석모형을 개발하였고, 건물의 영향을 고려할 수 있도록 구성하였다. 본 연구결과 지표류 유출 해석의 물리적 특성을 잘 반영하며, 도시지역의 복잡한 배수시스템 해석모형과 지표범람 모형을 통합한 모형 개발로 인해 더욱 정교한 도시지역에서의 홍수 범람 해석을 실시할 수 있을 것으로 판단된다. 본 모형의 개발로 침수상황의 시간별 진행과정을 분석함으로써 도시홍수에 대한 침수위험 지점 파악 및 주민대피지도 구축 등에 활용될 수 있을 것으로 판단된다. 있을 것으로 판단되었다.4일간의 기상변화가 자발성 기흉 발생에 영향을 미친다고 추론할 수 있었다. 향후 본 연구에서 추론된 기상변화와 기흉 발생과의 인과관계를 확인하고 좀 더 구체화하기 위한 연구가 필요할 것이다.게 이루어질 수 있을 것으로 기대된다.는 초과수익률이 상승하지만, 이후로는 감소하므로, 반전거래전략을 활용하는 경우 주식투자기간은 24개월이하의 중단기가 적합함을 발견하였다. 이상의 행태적 측면과 투자성과측면의 실증결과를 통하여 한국주식시장에 있어서 시장수익률을 평균적으로 초과할 수 있는 거래전략은 존재하므로 이러한 전략을 개발 및 활용할 수 있으며, 특히, 한국주식시장에 적합한 거래전략은 반전거래전략이고, 이 전략의 유용성은 투자자가 설정한 투자기간보다 더욱 긴 분석기간의 주식가격정보에 의하여 최대한 발휘될 수 있음을 확인하였다.(M1), 무역적자의

• Journal of Food Hygiene and Safety
• /
• v.27 no.2
• /
• pp.194-201
• /
• 2012

Facing a number of global food-related accidents, the concept and system for food traceability have been designed and introduced in many countries to manage the food-safety risks. To connect and harmonize the various food traceability-information in food traceability system according to the food supply chain, the coding system of identification-number for food-traceability has to be standardized. The GTIN (Global Trade Item Number) barcode system which has been globally standardized and implemented, is reviewed with the mandatory food-labeling regulation in expiration date of processed foods. The integration of GTIN-13 bar-code system for food-traceability is a crucial factor to expand its function in the food-related industrial areas. In this literature, the standard coding system of identification-number for food-traceability is proposed with 20 digit coding number which is combined with GTIN-13 bar-code (13 digit), expiration date (6 digit), and additional classification code (1 digit). This proposed standard coding system for identification-number has a several advantages in application for prohibiting the sale of hazard goods, food-recall, and inquiring food traceability-information. And also, this proposed coding system could enhance the food traceability system by communicating and harmonizing the information with the national network such as UNI-PASS and electronic Tax-invoice system. For the global application, the identification-number for food-traceability needs to be cooperated with the upcoming global standards such as GTIN-128 bar-code and GS1 DataBar.

• Journal of Intelligence and Information Systems
• /
• v.24 no.3
• /
• pp.89-117
• /
• 2018

The Korean companies are intensifying competition with not only domestic companies but also foreign companies in globalization. In this environment, it is essential activities not only for large companies but also Small and Medium Enterprises (SMEs) to get and develop the core competency. Particularly, SMEs that are inferior to resources of various aspects, such as financial resources etc., can make innovation through effective R&D investment. And then, SMEs can occupy a competency and can be survive at the environment. Conventionally, the method of "self-development" by using only the internal resources of the company has been dominant. Recently, however, R&D method through cooperation, also called "Open Innovation", is emerging. Especially SMEs are relatively short of available internal resources. Therefore, it is necessary to utilize technology and resources through cooperation with external companies(such as joint development or contract development etc.) rather than self-development R&D. In this context, we confirmed the effect of SMEs' factors on sales in Korea. Specifically, the factors that SMEs hold are classified as 'Technical characteristic', 'Company competency', and 'R&D activity' and analyzed how they influence the sales achieved as a result of R&D. The analysis was based on a two-year statistical survey conducted by the Korean government. In addition, we confirmed the influence of the factors on the sales according to the R&D method(Self-Development vs. Open Innovation), and also observed the influence change in 29 industrial categories. The results of the study are summarized as follows: First, regression analysis shows that twelve factors of SMEs have a significant effect on sales. Specifically, 15 factors included in the analysis, 12 factors excluding 3 factors were found to have significant influence. In the technical characteristic, 'imitation period' and 'product life cycle' of the technology were confirmed. In the company competency, 'R&D led person', 'researcher number', 'intellectual property registration status', 'number of R&D attempts', and 'ratio of success to trial' were confirmed. The R&D activity was found to have a significant impact on all included factors. Second, the influence of factors on the R&D method was confirmed, and the change was confirmed in four factors. In addition, these factors were found that have different effects on sales according to the R&D method. Specifically, 'researcher number', 'number of R&D attempts', 'performance compensation system', and 'R&D investment' were found to have significant moderate effects. In other words, the moderating effect of open innovation was confirmed for four factors. Third, on the industrial classification, it is confirmed that different factors have a significant influence on each industrial classification. At this point, it was confirmed that at least one factor, up to nine factors had a significant effect on the sales according to the industrial classification. Furthermore, different moderate effects have been confirmed in the industrial classification and R&D method. In the moderate effect, up to eight significant moderate effects were confirmed according to the industrial classification. In particular, 'R&D investment' and 'performance compensation system' were confirmed to be the most common moderating effect by each 12 times and 11 times in all industrial classification. This study provides the following suggestions: First, it is necessary for SMEs to determine the R&D method in consideration of the characteristics of the technology to be R&D as well as the enterprise competency and the R&D activity. In addition, there is a need to identify and concentrate on the factors that increase sales in R&D decisions, which are mainly affected by the industry classification to which the company belongs. Second, governments that support SMEs' R&D need to provide guidelines that are fit to their situation. It is necessary to differentiate the support for the company considering various factors such as technology and R&D purpose for their effective budget execution. Finally, based on the results of this study, we urge the need to reconsider the effectiveness of existing SME support policies.

• Journal of Intelligence and Information Systems
• /
• v.26 no.1
• /
• pp.23-45
• /
• 2020

Big data is creating in a wide variety of fields such as medical care, manufacturing, logistics, sales site, SNS, and the dataset characteristics are also diverse. In order to secure the competitiveness of companies, it is necessary to improve decision-making capacity using a classification algorithm. However, most of them do not have sufficient knowledge on what kind of classification algorithm is appropriate for a specific problem area. In other words, determining which classification algorithm is appropriate depending on the characteristics of the dataset was has been a task that required expertise and effort. This is because the relationship between the characteristics of datasets (called meta-features) and the performance of classification algorithms has not been fully understood. Moreover, there has been little research on meta-features reflecting the characteristics of multi-class. Therefore, the purpose of this study is to empirically analyze whether meta-features of multi-class datasets have a significant effect on the performance of classification algorithms. In this study, meta-features of multi-class datasets were identified into two factors, (the data structure and the data complexity,) and seven representative meta-features were selected. Among those, we included the Herfindahl-Hirschman Index (HHI), originally a market concentration measurement index, in the meta-features to replace IR(Imbalanced Ratio). Also, we developed a new index called Reverse ReLU Silhouette Score into the meta-feature set. Among the UCI Machine Learning Repository data, six representative datasets (Balance Scale, PageBlocks, Car Evaluation, User Knowledge-Modeling, Wine Quality(red), Contraceptive Method Choice) were selected. The class of each dataset was classified by using the classification algorithms (KNN, Logistic Regression, Nave Bayes, Random Forest, and SVM) selected in the study. For each dataset, we applied 10-fold cross validation method. 10% to 100% oversampling method is applied for each fold and meta-features of the dataset is measured. The meta-features selected are HHI, Number of Classes, Number of Features, Entropy, Reverse ReLU Silhouette Score, Nonlinearity of Linear Classifier, Hub Score. F1-score was selected as the dependent variable. As a result, the results of this study showed that the six meta-features including Reverse ReLU Silhouette Score and HHI proposed in this study have a significant effect on the classification performance. (1) The meta-features HHI proposed in this study was significant in the classification performance. (2) The number of variables has a significant effect on the classification performance, unlike the number of classes, but it has a positive effect. (3) The number of classes has a negative effect on the performance of classification. (4) Entropy has a significant effect on the performance of classification. (5) The Reverse ReLU Silhouette Score also significantly affects the classification performance at a significant level of 0.01. (6) The nonlinearity of linear classifiers has a significant negative effect on classification performance. In addition, the results of the analysis by the classification algorithms were also consistent. In the regression analysis by classification algorithm, Naïve Bayes algorithm does not have a significant effect on the number of variables unlike other classification algorithms. This study has two theoretical contributions: (1) two new meta-features (HHI, Reverse ReLU Silhouette score) was proved to be significant. (2) The effects of data characteristics on the performance of classification were investigated using meta-features. The practical contribution points (1) can be utilized in the development of classification algorithm recommendation system according to the characteristics of datasets. (2) Many data scientists are often testing by adjusting the parameters of the algorithm to find the optimal algorithm for the situation because the characteristics of the data are different. In this process, excessive waste of resources occurs due to hardware, cost, time, and manpower. This study is expected to be useful for machine learning, data mining researchers, practitioners, and machine learning-based system developers. The composition of this study consists of introduction, related research, research model, experiment, conclusion and discussion.