• Asia pacific journal of information systems
• /
• v.19 no.2
• /
• pp.157-178
• /
• 2009

Corporate credit rating is a very important factor in the market for corporate debt. Information concerning corporate operations is often disseminated to market participants through the changes in credit ratings that are published by professional rating agencies, such as Standard and Poor's (S&P) and Moody's Investor Service. Since these agencies generally require a large fee for the service, and the periodically provided ratings sometimes do not reflect the default risk of the company at the time, it may be advantageous for bond-market participants to be able to classify credit ratings before the agencies actually publish them. As a result, it is very important for companies (especially, financial companies) to develop a proper model of credit rating. From a technical perspective, the credit rating constitutes a typical, multiclass, classification problem because rating agencies generally have ten or more categories of ratings. For example, S&P's ratings range from AAA for the highest-quality bonds to D for the lowest-quality bonds. The professional rating agencies emphasize the importance of analysts' subjective judgments in the determination of credit ratings. However, in practice, a mathematical model that uses the financial variables of companies plays an important role in determining credit ratings, since it is convenient to apply and cost efficient. These financial variables include the ratios that represent a company's leverage status, liquidity status, and profitability status. Several statistical and artificial intelligence (AI) techniques have been applied as tools for predicting credit ratings. Among them, artificial neural networks are most prevalent in the area of finance because of their broad applicability to many business problems and their preeminent ability to adapt. However, artificial neural networks also have many defects, including the difficulty in determining the values of the control parameters and the number of processing elements in the layer as well as the risk of over-fitting. Of late, because of their robustness and high accuracy, support vector machines (SVMs) have become popular as a solution for problems with generating accurate prediction. An SVM's solution may be globally optimal because SVMs seek to minimize structural risk. On the other hand, artificial neural network models may tend to find locally optimal solutions because they seek to minimize empirical risk. In addition, no parameters need to be tuned in SVMs, barring the upper bound for non-separable cases in linear SVMs. Since SVMs were originally devised for binary classification, however they are not intrinsically geared for multiclass classifications as in credit ratings. Thus, researchers have tried to extend the original SVM to multiclass classification. Hitherto, a variety of techniques to extend standard SVMs to multiclass SVMs (MSVMs) has been proposed in the literature Only a few types of MSVM are, however, tested using prior studies that apply MSVMs to credit ratings studies. In this study, we examined six different techniques of MSVMs: (1) One-Against-One, (2) One-Against-AIL (3) DAGSVM, (4) ECOC, (5) Method of Weston and Watkins, and (6) Method of Crammer and Singer. In addition, we examined the prediction accuracy of some modified version of conventional MSVM techniques. To find the most appropriate technique of MSVMs for corporate bond rating, we applied all the techniques of MSVMs to a real-world case of credit rating in Korea. The best application is in corporate bond rating, which is the most frequently studied area of credit rating for specific debt issues or other financial obligations. For our study the research data were collected from National Information and Credit Evaluation, Inc., a major bond-rating company in Korea. The data set is comprised of the bond-ratings for the year 2002 and various financial variables for 1,295 companies from the manufacturing industry in Korea. We compared the results of these techniques with one another, and with those of traditional methods for credit ratings, such as multiple discriminant analysis (MDA), multinomial logistic regression (MLOGIT), and artificial neural networks (ANNs). As a result, we found that DAGSVM with an ordered list was the best approach for the prediction of bond rating. In addition, we found that the modified version of ECOC approach can yield higher prediction accuracy for the cases showing clear patterns.

• Journal of the Korean Society of Food Science and Nutrition
• /
• v.31 no.6
• /
• pp.1142-1149
• /
• 2002

This study was designed to standardize recipe and analyze the nutrients of Aguyjjim (stewed angler), a kind of native local foods in Busan. The results of this study are summarized as follows. The form of angler and its way to catch for a bait were recorded in Jasanobo, but the recipe was not written. While angler was used as manure or feed once, it was boomed for Aguyjjim since 80's, accredited as high-quality fish and protein-rich food. The standard recipe of Aguyjjim was prepared (through 3 times experimental cooking), being based on the recipe of cookbooks and restaurants. Then standard recipe was subjected to sensory evaluation. All sensory properties (appearance, smell, texture etc.) of Aguyiiim were evaluated as “like slight S, moderately”in hedonic scale. Energy per 100 g of the Aguyjjim was 89 kcal. It contained protein (10.7 g), fat (1.3 g), fiber (2.6 g), Ca (45 mg), Fe (1.3 mg) and so on. Amino acid is mainly comprised glutamic acid and aspartic acid. Free amino acid contents such as asparagine, glutamic acid, taurine were high. Fatty acid primarily consisted of unsaturated fatty acid like oleic acid, linoleic acid. Traditional foods are based on that of local foods. However, there has been little study looking into the recipes and nutritional value of local foods. Aiming at development of traditional food, the present study offered the standard cuisine of Aguyiiim that is a kind of local foods in Busan. In addition, it presented the composition of nutrients, amino acids and fatty acids. This data would be helpful for houses, restaurants and food service facilities to revive the local foods.

• Journal of the Korean Society of Food Science and Nutrition
• /
• v.37 no.11
• /
• pp.1523-1528
• /
• 2008

This study was carried out to compare the physicochemical characteristics and sensory quality of Sikhye (a Korean traditional rice beverage) prepared with specialty rice varieties. The results showed that Ilpum had higher hulling recovery, milled/brown rice ratio, and milling recovery than Sulgaeng, Baegjinju, Baegjinju 1, and Dongjinchal. The alkali digestive value, protein content, and amylose contents of Sulgaeng were 6.3, 7.3% and 19.3%, respectively. The highest brix degree was $10.00^{\circ}Bx$ in Baegjinju Sikhye. The turbidity appeared at 0.4440, 0.4100, 0.3828, 0.3372, 0.1414 in Ilpum, Baegjinju, Baegjinju 1, Sulgaeng, and Dongjinchal Sikhye, respectively. There were no significant differences in pH and maltose contents among the groups. The highest L-value was 44.62 in Ilpum Sikhye. The a-value and b-value were $-1.66{\sim}-0.70$ and $-9.18{\sim}-5.19$, respectively. Finally, the sensory evaluation results indicated that there were no significant differences in appearance, aroma, and taste between the groups, and the Sulgaeng Sikhye showed higher overall quality than the Dongjinchal Sikhye as the control.

• Journal of the Korean Society of Food Science and Nutrition
• /
• v.36 no.6
• /
• pp.785-793
• /
• 2007

This study was carried out to investigate the effects of red yeast rice addition on the physical characteristics of sourdough starters ($SD1{\sim}4$) and the quality characteristics of sourdough bread (CSB, $SDB1{\sim}3$). Moisture contents of red yeast rice, wheat flour, and rye flour were 6.15, 12.53, and 8.56%; ash contents were 0.15, 0.44, and 1.64%; protein contents were 7.30, 12.57, and 11.18%; crude lipid contents were 0.97, 1.16, and 2.49%, respectively. The pH decreased with increasing red yeast rice addition. Reducing sugar and total sugar increased with increasing red yeast rice addition. Lactic acid bacteria were not detected from 0 day for all samples and was maximum on the 2nd day. The yeast counts of SD1 were not detected from $0{\sim}4th$ day, but $SD2{\sim}4$ increased until the 3rd day. Peak time, peak value, and width of tail of CSB were higher than $SDB1{\sim}3$, and they increased with increasing red yeast rice contents of sourdough. The fermentation expansive power of $SDB1{\sim}3$ was higher than that of CSB. Baking loss and specific volume of SDB1 were higher than other samples and when baking loss of CSB and SDB1${\sim}$3 increased, the specific volume increased. L values decreased with increasing red yeast rice contents of sourdough bread whereas, a and b values increased. Springiness and brittleness of CSB and gumminess of SCB1 were lower than other samples. Springiness, brittleness, and gumminess increased with increasing red yeast rice content of sourdough bread. The sensory evaluation indicated that color, hardness, springiness, sweetness, and sourness increased with increasing red yeast rice content of sourdough bread. Aircell non-uniformity of SDB1 was lower than other samples, whereas off-flavor was higher than other samples.

• Journal of Korean Ophthalmic Optics Society
• /
• v.18 no.3
• /
• pp.347-356
• /
• 2013

Purpose: To evaluate the visual acuity and visual performance after implantation of a aspheric multifocal (ReSTOR$^{(R)}$ SN6AD3) intraocular lens (IOL). Methods: Nineteen cataract patients (30 eyes) implanted with an aspheric multifocal IOL (ReSTOR$^{(R)}$ SN6AD3) either unilaterally or bilaterally were participated. Visual acuity (VA) and objective optical performance were evaluated at the time of preoperation, 1 week, 1 month, and 3 month after operation. At 3 month of post-operation, objective visual performance were measured and compared with the 38 eyes of 20 age-matched normal control. Distance VA was measured by using the ETDRS LCD chart and intermediate and near visual acuity were measured using Jaeger chart. Objective visual performance was assessed preoperatively and 1 week, 1 month and 3 month postoperatively using a double-pass system (Optical Quality Analysis System) with a 4-mm pupil diameter, the OSI (objective scatter index), MTF (modulation transfer function) cut off and strehl ratio. At 3 month of post-operation, visual acuity and visual performance compared with age matched normal control. Results: The uncorrected distance VA, OSI, MTF cut off and strehl ratio were significantly improved (p<0.05) until 1 month postoperatively. Visual performance of MTF cut off and strehl ratio after 3 month of operation were significantly improved compared to the normal control (p=0.063, p=0.103 respectively), however, OSI was higher than normal control. Patients implanted with aspheric multifocal IOL were satisfied with distance and near VA however, were unsatisfied with intermediate VA and reported glare and halos. Conclusions: The visual performance reaches to a stable condition in 1 month of implantation of aspheric multifocal IOL and improved to the level of age-mated normal patients. Also patients were satisfied with their quality of vision, however, intermediate VA, glare and halos were reported as complications.

• Journal of the Korean Society of Food Science and Nutrition
• /
• v.35 no.2
• /
• pp.199-204
• /
• 2006

This study was conducted to investigate the quality characteristics of noodles prepared with the addition of nanofiltered (NF) powder of sunmul. Noodles were prepared with different levels $(0\%,\;1.5\%,\;3\%\;and\;5\%,\;w/w)$ of NF powder and physico-chemical properties were examined. Results of rapid visco analyzer showed that peak, trough, final viscosity and set back decreased as the NF powder level increased. The weight and volume of cooked noodles increased with the addition of NF powder. Turbidity of soup also increased as the amount of NF powder increased, indicating higher cooking loss. The color of wet and cooked noodles became greenish yellow as the NF powder level increased. Hardness, springiness, gumminess and brittleness of cooked noodles decreased with the increasing amount of NF powder. Results of sensory evaluation showed that noodles prepared with up to $3\%$ addition of NF powder was considered to be as acceptable as noodles prepared without NF powder.

• Journal of Korean Society of Environmental Engineers
• /
• v.29 no.4
• /
• pp.404-411
• /
• 2007

Adsorption performance of odorous compounds such as geosmin and 2-MIB on granular activated carbon were evaluated in this study. The coal-based activated carbon was found more effective than other carbons in adsorption of geosmin and 2-MIB. The wood-based virgin activated carbon was less effective than coconut- and coal-based carbon in adsorption nevertheless having larger pore volume and specific surface area than others carbons. The maximum adsorption capacity(X/M) of coal-based activated carbon for geosmin and 2-MIB was $1.2\sim1.9$ and $2.1\sim2.6$ times larger than coconut- and wood-based virgin activated carbon, respectively. Carbon usage rate (CUR) of coal-, coconut- and wood-based virgin activated carbons for geosmin and 2-MIB were 1.72 and 1.44 g/day, 1.72 and 2.05 g/day and 2.12 and 1.90 g/day, respectively. In the evaluation of adsorption isotherm of geosmin and 2-MIB for coal-, coconut- and wood-based virgin activated carbons, k value of 2-MIB was lower than geosmin, It menas 2-MIB is more difficult to remove by activated carbon adsorption than geosmin. The relationship of max. adsorption versus total pore volume of coconut- and wood-based virgin and used activated carbon for geosmin and 2-MIB were $y=264,459\times-79,047(R^2=0.95)$, $y=319,650\times-101,762(R^2=0.93)$.

• Journal of Animal Science and Technology
• /
• v.49 no.6
• /
• pp.871-880
• /
• 2007

This study was conducted to investigate the relationship between the socio-demographic factors and the Korean consumers’ palatability evaluation and to collect the sensory information for development of prediction palatability model of Hanwoo beef. Ten cuts〔Abjin (short plate), Bosup (top sirloin), Cheggt (striploin), Dngsim (loin), Guri (chuck tender), Hongduke (eye of round), Moksim (chuck roll), Sulgit (bottom round), Udoon (top round), Yangi (brisket)〕 were separated from 10 Hanwoo bulls beef(～24 months old) and prepared with the same manner for 3 different cooking methods such as boiling, grilling and roasting. The cooked beef samples were served to 650 consumers recruited from Seoul, Joongbu, Honam and Youngnam locations and evaluated tenderness, juiciness, flavor and overall acceptability. The living location, age, gender, occupation, eating habit and cut were significantly related with the sensory properties when bull beef were cooked as boiled cooking(p<0.05). In grill cooking, only occupation, income and cut had the significant effect on the overall acceptability of Hanwoo bull beef(p<0.01). When bull beef were cooked as Korean traditional roast cooking, consumer’s sensory properties were significantly different by the living location, age, occupation, income, eating habit of consumers and cut(p<0.05). Also, results from the principal component analysis showed that palatability scores of Korean consumers were decided with different contribution scores of tenderness, juiciness, flavor and overall acceptability of cut depending on cooking methods. In conclusion, Korean consumers’ palatability for Hanwoo bull beef were related with the socio-demographic factors and the sensory results were different by cut and cooking methods.

• Journal of Animal Science and Technology
• /
• v.49 no.6
• /
• pp.761-772
• /
• 2007

This study was conducted to evaluate effects of dietary pine cone meal on growth performance, blood characteristics, carcass quality and fecal noxious gases compounds in finishing pigs. The total of sixty [(Landrace×Yorkshire)×Duroc] pigs(86.01±0.25kg in average initial body weight) were used in 35 days assay. Dietary treatments included 1) T1(2% cottonwood sawdust), 2) T2(1% cottonwood sawdust+1% pine cone meal) and 3) T3(2% pine cone meal). There were three dietary treatments with five replicate pens per treatment and four pigs per pen. During the overall periods, there were no significant differences in ADG(Average daily gain), ADFI(Average daily feed intake) and gain/feed ratio among treatments (P>0.05). Also, Nutrient digestibilities and blood characteristics were not affected by dietary treatments. At the end of this experiment, a*-value of logissimus dorsi muscle color and sensory evaluation color were higher in T3 treatment than T1 treatment(P<0.05). In fatty acid contents of lean, C18:1 and total MUFA were significantly lower in T1 treatment than other treatments(P<0.05). However, total ω6 and total PUFA were higher in T1 treatment than T2 treatment(P<0.05). In fatty acid contents of fats, total SFA was significantly higher in T2 treatment than T3 treatment(P<0.05). C18:1 was higher in T2 treatment than T1 treatment(P<0.05). There were no significant differences in fecal noxious gas compounds among the treatments. In conclusion, the results of the experiment was affected by dietary supplementation of pine cone meal on meat color and fatty acid composition of pork in finishing pigs.

• Journal of the Korean Society of Radiology
• /
• v.14 no.2
• /
• pp.129-138
• /
• 2020

Due to the nature of the Rosenberg Method, the patient needs to maintain posture for a certain period of time, and the joint space is observed in various forms depending on the angle of knee flexion, which causes difficulties in examination. In order to solve these problems, Image quality was evaluated in order to evaluate the usefulness of the assistive device by making the assistive device itself. SNR and CNR analysis about the presence or absence of an assistive device using the extremity phantom and the angle of the assistive device itself were not statistically significant(p < 0.05). As a result of measuring the distance between the right and left edges of the medial condyle based on the presence or absence of an assist device, and the absence of assist device (96.00±40.6 mm) and presence of an assist device (134.86±17.68 mm) were statistically significant (p <0.05). To find the aLDFA relationship about the femur and tibia, we measured the right and left aLDFA based on the presence or absence of assist device. As a result, the absence of the right-side aLDFA assist device (74.63°±4.87) and the presence of assist device (79.64°±3.65) were statistically significant (p <0.05). The absence of the left-side aLDFA assist device (76.39°±4.62) and the presence of assist device (79.64°±3.65) were statistically significant (p < 0.05). but, As a result of measuring the distance of the overlapping parts of the right and left proximal tibiofibular joint and the lateral condyle, There were no statistically significant differences between the right and left sides. In conclusion, we confirmed that we can obtain Diagnostically valuable images with a constant knee-to-knee spacing using an assist device we self-created. In addition, we could learn through aLDFA relationship between femur and tibial that the smaller the angle, the more medial condyle overlaps with JSW, We also confirmed the significance by deriving similar values on the normal range values of aLDFA using assist devices. However, it is considered necessary to pay attention to internal and external rotations in order to obtain good quality images by evaluating the distance of overlapping parts between proximal tibiofibular joint and lateral condyle.