• 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 Intelligence and Information Systems
• /
• v.21 no.4
• /
• pp.111-131
• /
• 2015

There are only a handful number of research conducted on pattern analysis of corporate distress as compared with research for bankruptcy prediction. The few that exists mainly focus on audited firms because financial data collection is easier for these firms. But in reality, corporate financial distress is a far more common and critical phenomenon for non-audited firms which are mainly comprised of small and medium sized firms. The purpose of this paper is to classify non-audited firms under distress according to their financial ratio using data mining; Self-Organizing Map (SOM). SOM is a type of artificial neural network that is trained using unsupervised learning to produce a lower dimensional discretized representation of the input space of the training samples, called a map. SOM is different from other artificial neural networks as it applies competitive learning as opposed to error-correction learning such as backpropagation with gradient descent, and in the sense that it uses a neighborhood function to preserve the topological properties of the input space. It is one of the popular and successful clustering algorithm. In this study, we classify types of financial distress firms, specially, non-audited firms. In the empirical test, we collect 10 financial ratios of 100 non-audited firms under distress in 2004 for the previous two years (2002 and 2003). Using these financial ratios and the SOM algorithm, five distinct patterns were distinguished. In pattern 1, financial distress was very serious in almost all financial ratios. 12% of the firms are included in these patterns. In pattern 2, financial distress was weak in almost financial ratios. 14% of the firms are included in pattern 2. In pattern 3, growth ratio was the worst among all patterns. It is speculated that the firms of this pattern may be under distress due to severe competition in their industries. Approximately 30% of the firms fell into this group. In pattern 4, the growth ratio was higher than any other pattern but the cash ratio and profitability ratio were not at the level of the growth ratio. It is concluded that the firms of this pattern were under distress in pursuit of expanding their business. About 25% of the firms were in this pattern. Last, pattern 5 encompassed very solvent firms. Perhaps firms of this pattern were distressed due to a bad short-term strategic decision or due to problems with the enterpriser of the firms. Approximately 18% of the firms were under this pattern. This study has the academic and empirical contribution. In the perspectives of the academic contribution, non-audited companies that tend to be easily bankrupt and have the unstructured or easily manipulated financial data are classified by the data mining technology (Self-Organizing Map) rather than big sized audited firms that have the well prepared and reliable financial data. In the perspectives of the empirical one, even though the financial data of the non-audited firms are conducted to analyze, it is useful for find out the first order symptom of financial distress, which makes us to forecast the prediction of bankruptcy of the firms and to manage the early warning and alert signal. These are the academic and empirical contribution of this study. The limitation of this research is to analyze only 100 corporates due to the difficulty of collecting the financial data of the non-audited firms, which make us to be hard to proceed to the analysis by the category or size difference. Also, non-financial qualitative data is crucial for the analysis of bankruptcy. Thus, the non-financial qualitative factor is taken into account for the next study. This study sheds some light on the non-audited small and medium sized firms' distress prediction in the future.

• Korean Journal of Weed Science
• /
• v.6 no.2
• /
• pp.174-190
• /
• 1986

A series of laboratory and greenhouse experiments were conducted to find out the possibility of tissue culture and cell culture methods as a tool to study herbicidal behaviour and herbicide tolerance screening from 1985 to 1986 at the Yeongnam Crop Experiment Station. For dehulled-rice culture, pure agar medium was the most appropriate in rice growth campared to other media used for plant tissue culture method. All the media but the pure agar medium resulted in growth retardance by approximately 50% and this effect was more pronounced to root growth than shoot growth. Herbicidal phytotoxicity was enhanced under light condition for butachlor, 2.4-D, and propanil while this effect was reversed for DPX F-5384 and CGA 142464, respectively. And also, herbicides of butachlor, chlornitrofen, oxadiazon, and BAS-514 resulted in more phytotoxic effect when shoot and root of rice were exposed to herbicide than root exposure only while other used herbicides exhibited no significant difference between two exposure regimes. Similar response was obtained from Echinochloa crusgalli even though the degree of growth retardance was much greater. Particularly, butachlor, 2.4-D, chlornitrofen, oxadiaxon, pyrazolate and BAS-514 totally inhibited chlorophyll biosynthesis even at the single contact of root. Apparent cultivar differences to herbicide were observed at the young seedling culture method and dehulled rice cultivars were more tolerant in DPX F-5384, NC-311, pyrazolate and pyrazoxyfen, respectively. For derant than other types or rice cultivar in butachlor, pretilachlor, perfluidone and oxadiazon while Tongil-type rice cultivars were more tolerant in DPXF-5384, NC-311, Pyrazolate and Pyrazoxyfen, respectively. For dehulled rice culture, on the other hand, Japonica-type rice cultivar was less tolerant to herbicides of butachlor, propanil, chlornitrofen and oxadiazon that was reversed trend to young seedling culture test. Cultivar differences were also exhibited within same cultivar type. In general, relatively higher tolerant cultivars were Milyang 42, Cheongcheongbyeo, Samgangbyeo, Chilseoungbyeo for Tongil-type, Somjinbyeo for Japonica-type and IR50 for Indica-type, respectively. The response of callus growth showed similar to dehulled rice culture method in all herbicides regardless of property variables. However, concentration response was much sensitive in callus response. The concentration ranges of $10^{-9}M-10^(-8)M$ were appropriate to distinguish the difference between herbicides for E. crusgalli callus growth. Among used herbicides, BAS-514 was the most effective to E. crusgalli callus growth. Based on the above results, tissue culture method could be successfully used as a tool for studying herbicidal behaviour and tolerance screening to herbicide.

• Journal of Intelligence and Information Systems
• /
• v.17 no.1
• /
• pp.53-69
• /
• 2011

Vision and voice-based technologies are commonly utilized for human-robot interaction. But it is widely recognized that the performance of vision and voice-based interaction systems is deteriorated by a large margin in the real-world situations due to environmental and user variances. Human users need to be very cooperative to get reasonable performance, which significantly limits the usability of the vision and voice-based human-robot interaction technologies. As a result, touch screens are still the major medium of human-robot interaction for the real-world applications. To empower the usability of robots for various services, alternative interaction technologies should be developed to complement the problems of vision and voice-based technologies. In this paper, we propose the use of accelerometer-based gesture interface as one of the alternative technologies, because accelerometers are effective in detecting the movements of human body, while their performance is not limited by environmental contexts such as lighting conditions or camera's field-of-view. Moreover, accelerometers are widely available nowadays in many mobile devices. We tackle the problem of classifying acceleration signal patterns of 26 English alphabets, which is one of the essential repertoires for the realization of education services based on robots. Recognizing 26 English handwriting patterns based on accelerometers is a very difficult task to take over because of its large scale of pattern classes and the complexity of each pattern. The most difficult problem that has been undertaken which is similar to our problem was recognizing acceleration signal patterns of 10 handwritten digits. Most previous studies dealt with pattern sets of 8~10 simple and easily distinguishable gestures that are useful for controlling home appliances, computer applications, robots etc. Good features are essential for the success of pattern recognition. To promote the discriminative power upon complex English alphabet patterns, we extracted 'motion trajectories' out of input acceleration signal and used them as the main feature. Investigative experiments showed that classifiers based on trajectory performed 3%~5% better than those with raw features e.g. acceleration signal itself or statistical figures. To minimize the distortion of trajectories, we applied a simple but effective set of smoothing filters and band-pass filters. It is well known that acceleration patterns for the same gesture is very different among different performers. To tackle the problem, online incremental learning is applied for our system to make it adaptive to the users' distinctive motion properties. Our system is based on instance-based learning (IBL) where each training sample is memorized as a reference pattern. Brute-force incremental learning in IBL continuously accumulates reference patterns, which is a problem because it not only slows down the classification but also downgrades the recall performance. Regarding the latter phenomenon, we observed a tendency that as the number of reference patterns grows, some reference patterns contribute more to the false positive classification. Thus, we devised an algorithm for optimizing the reference pattern set based on the positive and negative contribution of each reference pattern. The algorithm is performed periodically to remove reference patterns that have a very low positive contribution or a high negative contribution. Experiments were performed on 6500 gesture patterns collected from 50 adults of 30~50 years old. Each alphabet was performed 5 times per participant using $Nintendo{(R)}$ $Wii^{TM}$ remote. Acceleration signal was sampled in 100hz on 3 axes. Mean recall rate for all the alphabets was 95.48%. Some alphabets recorded very low recall rate and exhibited very high pairwise confusion rate. Major confusion pairs are D(88%) and P(74%), I(81%) and U(75%), N(88%) and W(100%). Though W was recalled perfectly, it contributed much to the false positive classification of N. By comparison with major previous results from VTT (96% for 8 control gestures), CMU (97% for 10 control gestures) and Samsung Electronics(97% for 10 digits and a control gesture), we could find that the performance of our system is superior regarding the number of pattern classes and the complexity of patterns. Using our gesture interaction system, we conducted 2 case studies of robot-based edutainment services. The services were implemented on various robot platforms and mobile devices including $iPhone^{TM}$. The participating children exhibited improved concentration and active reaction on the service with our gesture interface. To prove the effectiveness of our gesture interface, a test was taken by the children after experiencing an English teaching service. The test result showed that those who played with the gesture interface-based robot content marked 10% better score than those with conventional teaching. We conclude that the accelerometer-based gesture interface is a promising technology for flourishing real-world robot-based services and content by complementing the limits of today's conventional interfaces e.g. touch screen, vision and voice.

• Journal of Distribution Science
• /
• v.9 no.2
• /
• pp.111-122
• /
• 2011

The purpose of this study is to survey the successful cases of small and medium Business Systematization Cognition by examining their entrepreneurial characteristics and analysing the factors affecting their success. To that end, previous studies on the association types of small businesses were studied. A research model was developed, and research hypotheses for an empirical analysis were established upon it. Suh et al. (2010) insist on the importance of Small Business Systematization in Korea but also show that small business performance is suffering: they are too small to stand alone. That is why association is so crucial for them: they must stand together. Unfortunately, association is difficult, as they have few specific links and little motivation. Even in franchising networks, association tends to be initiated by big franchisers, not small ones. In that sense, association among small businesses is crucial for their long-term survival. With this in mind, this study examines how they think and feel about the issue of 'Industrial Classification', how important Industrial Classification is to their business success, and what kinds of problems it raises in the markets. This study seeks the different cognitions among the association types of small businesses from the perspectives of participation motivation, systematization expectation, policy demand level, and management performance. We assume that different industrial classification types of small businesses will have different cognitions concerning these factors. There are four basic industrial classification types of small businesses: retail sales, restaurant, service, and manufacturing. To date, most of the studies in this area have focused on collecting data on the external environments of small businesses or performing statistical analyses on their status. In this study, we surveyed 4 market areas in Busan, Masan, and Changwon in Korea, where business associations consist of merchants, shop owners, and traders. We surveyed 330 shops and merchants by sending a questionnaire or visiting. Finally, 268 questionnaires were collected and used for the analysis. An ANOVA, T-test, and regression analyses were conducted to test the research hypotheses. The results demonstrate that there are differences in cognition depending upon the industrial classification type. Restaurants generally have a higher cognition concerning job offer problems and a lower cognition concerning their competitiveness. Restaurants also depend more on systematization expectation than do the other industrial classification types. On the policy demand level, restaurants have a higher cognition. This study identifies several factors that are contributing to management performance through differences in cognition that depend upon association type: systematization expectation and policy demand level have positive effects on management performance; participation motivation has a negative effect on management performance. We confirm also that the image factors of different cognitions are linked to an awareness of the value of systematization and that these factors show sequential and continual patterns in the course of generating performances. In conclusion, this study carries significant implications in its classifying of small businesses into the four different associational types (retail sales, restaurant, services, and manufacturing). We believe our study to be the first one to conduct an empirical survey in this subject area. More studies in this area will likely use our research frameworks. The data show that regionally based industrial classification associations such as those in rural cities or less developed areas tend to suffer more problems than those in urban areas. Moreover, restaurants suffer more problems than the norm. Most of the problems raised in this study concern the act of 'associating itself'. Most associations have serious difficulties in associating. On the other hand, the area where they have the least policy demand is that of service types. This study contributes to the argument that associating, rather than financial assistance or management consulting, promotes the start-up and managerial performance of small businesses. This study also has some limitations. The main limitation is the number of questionnaires. We could not survey all the industrial classification types across the country because of budget and time limitations. If we had, we could have produced many more useful results and enhanced the precision of our analysis. The history of systemization is very short and the number of industrial classification associations is relatively low in Korea. We should keep in mind, though, that this is very crucial to systemization entrepreneurs starting their businesses, as it can heavily affect their chances of success. Being strongly associated with each other might be critical to the business success of industrial classification members. Thus, the government needs to put more effort and resources into supporting the drive of industrial classification members to become more strongly associated.

• FLOWER RESEARCH JOURNAL
• /
• v.18 no.3
• /
• pp.186-192
• /
• 2010

An infrared heating system, installed in a small venlo-type glasshouse ($280m^2$) in Gyeongsang National University, Jinju, Korea, was used to investigate its heating effect with potted Phalaenopsis, Schefflera arboricola 'Hongkong', Ficus elastica 'Variegata', and Rosa hybrida 'Yellow King' as the test plants. Temperature changes in test plants with the system turned 'On' and 'Off' were measured by using an infrared camera and the consumption of electricity by this infrared heating system was measured and analyzed. In potted Phalaenopsis, when the set air temperature of the greenhouse was $18^{\circ}C$, temperature of leaves and the growing medium were $22.8{\sim}27^{\circ}C$ and $21.3{\sim}24.3^{\circ}C$, respectively. In such tall plants as Schefflera arboricola 'Hongkong' and Ficus elastica 'Variegata', the upper part showed the highest temperature of 24.0 and $26.9^{\circ}C$, respectively. From the results of temperature change measurements, the plant temperatures were near or above the set point temperatures with some fluctuations depending on the position or distance from the infrared heating system. When air temperature between night and dawn dropped sharply, plant temperatures were maintained close to the set temperature ($18^{\circ}C$). There was a significant difference between 'On' and 'Off' states of the infrared heating system in average temperatures of root zone and leaf: 21.8 and $17.8^{\circ}C$ with the system 'On' and 20.4 and $15.5^{\circ}C$ with the system 'Off', respectively, in a cut rose Rosa hybrida 'Yellow King'. The heating load was about $24,850{\sim}35,830kcal{\cdot}h^{-1}$, which comes to about 27,000~40,000 won in Korean currency when calculated in terms of the cost of heating by a hot water heating system heated by petroleum. The cost for heating by the infrared heating system was about 35% of that of a hot water heating system. With the infrared heating system, the air temperature during the night was maintained slightly lower than the set point air temperature, probably due to the lack of air tightness of the glasshouse. Therefore, glasshouses with an infrared heating system requires further investigation including the installation space of the heat-emitting units, temperature sensor positions, and convection.

• Korean Journal of Soil Science and Fertilizer
• /
• v.19 no.2
• /
• pp.157-165
• /
• 1986

This experiment was conducted to find out the some physiological characteristics and nitrogen fixation activities under free-living conditions of indigenous rhizobia isolated from soybean-cultivated (30 fields) and un-cultivated (30 fields) soil. The results were summerized as follows: 1. In free-living condition, only 12.8% and 6.4% of the indigenous rhizobia isolated from soybean cultivated (133 strains) and un-cultivated (125 strains) soils were nitrogenase positive as more than 4 n mole $C_2H_4$ per tube per hour by acetylene reduction assay. 2. The acid-producing rhizobia on litmus milk test was observed to be 20% of the total 160 strains isolated from soybean cultivated soil but about 34% of 166 strains isolated from un-cultivated soil. And the serum zone positive strains were higher in the soybean un-cultivated soil than cultivated soil. 3. The population ratio of fast-to slow-growing indigenous rhizobia based on growth pattern of AMA medium was 35.6% to 64.4% of the total 346 strains. 4. The population of indigenous Rhizobium japonicum counted by MPN method was ranged from $9.2{\times}10^2$ cells per gram of soil in soybean un-cultivated soil to $2.3{\times}10^4$ cells per gram of soil soybean cultivated soil. The number of indigenous R. japonicum in 0-10cm depth of surface layer was higher than low layer.

• Korean Journal of Food Science and Technology
• /
• v.40 no.4
• /
• pp.430-435
• /
• 2008

The sensory characteristics of imported (two dry, two sweet, and one medium dry wines) and domestic (one sweet wine) red wine were evaluated by 250 panels. The preferences of aroma, color, sweetness, tartness, astringency, and overall acceptability were determined by 5-point just-about-right scale. Among six wines, B sample obtained the highest mean overall acceptability score of 3.67 and its chemical and sensory characteristics were as follow: cherry or strawberry aroma, 9.4 brix, 3.7% of sugar content, pH 3.5, 10% of ethanol, 0.14% of tannin, 5.74 mg/mL of total organic acids, and color of $L(12.04{\pm}0.01)$, $a(33.90{\pm}0.19)$, and $b(8.22{\pm}0.00)$. These results revealed that Korean consumers generally prefer sweet taste to dry one, flat taste to bitter one, and fruity aroma to others. Panels preferred a red wine containing high sugar content of $5{\sim}10%$. On the other hand, panels showed low preference to the wine containing tannin above 0.16%. However, these general trends were slightly varied depending on their ages and consuming frequencies.

• Horticultural Science & Technology
• /
• v.33 no.6
• /
• pp.891-899
• /
• 2015

The lily symptomless virus (LSV) is the most common virus in Korean native lilies and causes various types of damage to overall plant growth. This study was carried out to investigate the elimination rate of the LSV by the in vitro scale culture (scaling) method in Korean native lilies and to test reinfection rates of the LSV under several field culture conditions of bulb production. Four Korean native lilies (Lilium dauricum, L. distichum, L. lancifolium, and L. maximowitzii) were used and their scales were cultured in vitro for micro-scale formation. The micro-scales were subcultured repeatedly using MS culture medium supplemented with 30 or $90g{\cdot}L^{-1}$ sucrose. The culture conditions were $24{\mu}mol{\cdot}m^{-2}{\cdot}s^{-1}$ PPFD with 16 hour daylength using fluorescent lamps and maintained at $22{\pm}2^{\circ}C$. The virus-free bulblets were grown for one to three years in the greenhouse and transplanted to the field in October or March. Virus infection rates were investigated by direct tissue blotting immunobinding assays and measurement of chlorophyll and protein contents. Virus-free plants could be obtained from the 5th subculture of micro-scales in L. lancifolium and L. maximowitzii or from primary culture in L. dauricum and L. distichum. LSV-free plants were reinfected during bulb production in the field. Reinfection rates were higher at older bulb ages and under higher planting density. The plants planted in October and at inland Gyeongsan had higher infection rates than those planted in March and at coastal area Pohang. The reinfection rate of L. maximowitzii was higher than those of L. dauricum and L. lancifolium. The LSV-infected plants had lower chlorophyll contents and unchanged protein contents compared to virus-free plants.

• Journal of Animal Environmental Science
• /
• v.6 no.3
• /
• pp.175-184
• /
• 2000

A lab-scale MF membrane reactor was installed to investigate the membrane permeability, characteristics of membrane fouling at each conditions, and quality of permeate (liquid livestock manure) in the separation of solid-matters using membrane. Experiment was divided three filtration type such as follows; continuous filtration, gravity filtration, and intermittent filtration. As a result of experiment, flux 1 LMH was maintained for 7days, and trans-membrane pressure(TMP) was increased gradually under 10cmHg, but it was increased immediately after 10cmHg, respectively. However, the flux was increased, the Tmax was decreased exponential more and more. During the pure-flux test, most of the fouling of membrane was reversible. At the gravity filtration, permeate could be obtained as 1.75 LMH for 3.5days without any other electronic pressure. As an investigation of membrane surface, this study could be decided that the reason of fouling at the lower flux (Run 1 and 2) was attached matters in membrane surface, but at the higher flux (Run 4-6) was concentration polarization.