• Journal of Intelligence and Information Systems
• /
• v.18 no.2
• /
• pp.29-45
• /
• 2012

Bond rating is regarded as an important event for measuring financial risk of companies and for determining the investment returns of investors. As a result, it has been a popular research topic for researchers to predict companies' credit ratings by applying statistical and machine learning techniques. The statistical techniques, including multiple regression, multiple discriminant analysis (MDA), logistic models (LOGIT), and probit analysis, have been traditionally used in bond rating. However, one major drawback is that it should be based on strict assumptions. Such strict assumptions include linearity, normality, independence among predictor variables and pre-existing functional forms relating the criterion variablesand the predictor variables. Those strict assumptions of traditional statistics have limited their application to the real world. Machine learning techniques also used in bond rating prediction models include decision trees (DT), neural networks (NN), and Support Vector Machine (SVM). Especially, SVM is recognized as a new and promising classification and regression analysis method. SVM learns a separating hyperplane that can maximize the margin between two categories. SVM is simple enough to be analyzed mathematical, and leads to high performance in practical applications. SVM implements the structuralrisk minimization principle and searches to minimize an upper bound of the generalization error. In addition, the solution of SVM may be a global optimum and thus, overfitting is unlikely to occur with SVM. In addition, SVM does not require too many data sample for training since it builds prediction models by only using some representative sample near the boundaries called support vectors. A number of experimental researches have indicated that SVM has been successfully applied in a variety of pattern recognition fields. However, there are three major drawbacks that can be potential causes for degrading SVM's performance. First, SVM is originally proposed for solving binary-class classification problems. Methods for combining SVMs for multi-class classification such as One-Against-One, One-Against-All have been proposed, but they do not improve the performance in multi-class classification problem as much as SVM for binary-class classification. Second, approximation algorithms (e.g. decomposition methods, sequential minimal optimization algorithm) could be used for effective multi-class computation to reduce computation time, but it could deteriorate classification performance. Third, the difficulty in multi-class prediction problems is in data imbalance problem that can occur when the number of instances in one class greatly outnumbers the number of instances in the other class. Such data sets often cause a default classifier to be built due to skewed boundary and thus the reduction in the classification accuracy of such a classifier. SVM ensemble learning is one of machine learning methods to cope with the above drawbacks. Ensemble learning is a method for improving the performance of classification and prediction algorithms. AdaBoost is one of the widely used ensemble learning techniques. It constructs a composite classifier by sequentially training classifiers while increasing weight on the misclassified observations through iterations. The observations that are incorrectly predicted by previous classifiers are chosen more often than examples that are correctly predicted. Thus Boosting attempts to produce new classifiers that are better able to predict examples for which the current ensemble's performance is poor. In this way, it can reinforce the training of the misclassified observations of the minority class. This paper proposes a multiclass Geometric Mean-based Boosting (MGM-Boost) to resolve multiclass prediction problem. Since MGM-Boost introduces the notion of geometric mean into AdaBoost, it can perform learning process considering the geometric mean-based accuracy and errors of multiclass. This study applies MGM-Boost to the real-world bond rating case for Korean companies to examine the feasibility of MGM-Boost. 10-fold cross validations for threetimes with different random seeds are performed in order to ensure that the comparison among three different classifiers does not happen by chance. For each of 10-fold cross validation, the entire data set is first partitioned into tenequal-sized sets, and then each set is in turn used as the test set while the classifier trains on the other nine sets. That is, cross-validated folds have been tested independently of each algorithm. Through these steps, we have obtained the results for classifiers on each of the 30 experiments. In the comparison of arithmetic mean-based prediction accuracy between individual classifiers, MGM-Boost (52.95%) shows higher prediction accuracy than both AdaBoost (51.69%) and SVM (49.47%). MGM-Boost (28.12%) also shows the higher prediction accuracy than AdaBoost (24.65%) and SVM (15.42%)in terms of geometric mean-based prediction accuracy. T-test is used to examine whether the performance of each classifiers for 30 folds is significantly different. The results indicate that performance of MGM-Boost is significantly different from AdaBoost and SVM classifiers at 1% level. These results mean that MGM-Boost can provide robust and stable solutions to multi-classproblems such as bond rating.

• The Journal of Advanced Prosthodontics
• /
• v.8 no.2
• /
• pp.101-109
• /
• 2016

PURPOSE. To investigate the void parameters within the resin cements used for fiber post cementation by micro-CT (${\mu}CT$) and regional push-out bonding strength. MATERIALS AND METHODS. Twenty-one, single and round shaped roots were enlarged with a low-speed drill following by endodontic treatment. The roots were divided into three groups (n=7) and fiber posts were cemented with Maxcem Elite, Multilink N and Superbond C&B resin cements. Specimens were scanned using ${\mu}CT$ scanner at resolution of $13.7{\mu}m$. The number, area, and volume of voids between dentin and post were evaluated. A method of analysis based on the post segmentation was used, and coronal, middle and apical thirds considered separately. After the ${\mu}CT$ analysis, roots were embedded in epoxy resin and sectioned into 2 mm thick slices (63 sections in total). Push-out testing was performed with universal testing device at 0.5 mm/min cross-head speed. Data were analyzed with Kruskal-Wallis and Mann-Whitney U tests (${\alpha}=.05$). RESULTS. Overall, significant differences between the resin cements and the post level were observed in the void number, area, and volume (P<.05). Super-Bond C&B showed the most void formation ($44.86{\pm}22.71$). Multilink N showed the least void surface ($3.51{\pm}2.24mm^2$) and volume ($0.01{\pm}0.01mm^3$). Regional push-out bond strength of the cements was not different (P>.05). CONCLUSION. ${\mu}CT$ proved to be a powerful non-destructive 3D analysis tool for visualizing the void parameters. Multilink N had the lowest void parameters. When efficiency of all cements was evaluated, direct relationship between the post region and push-out bonding strength was not observed.

• Journal of Industrial Convergence
• /
• v.19 no.5
• /
• pp.21-34
• /
• 2021

The purpose of this study is to find out what the investment efficiency of BW is from an investor's point of view and to suggest an efficient investment plan to investors. The research method is to investigate the coupon interest rate, maturity interest rate, issuance date, right exercise start and end date, maturity date, exercise price, etc. for BW issued from 2014 to July 2021. By connecting them, it was attempted to quantitatively understand the efficiency of investment in BW and the effect of new stock acquisitions. As a result of the study, the ratio of the number of days in excess of the exercise price was 41.3% of the available days for new stocks, so it was analyzed that the investment efficiency of bonds with warrants was not high. The return on the exercise start date was 24.8% on average and the return on the end date was 52.6% on average, showing a positive return on average, so it was derived in line with investor expectations. The number of stocks with negative returns on the exercise start date was 1.47 times higher than the number of stocks with positive returns, and the number of stocks with negative returns on the end date was 1.16 times higher than the number of positive stocks.

• Proceedings of the KSME Conference
• /
• 2007.05b
• /
• pp.2971-2976
• /
• 2007

The mode change from Taylor cone-jet to dripping in electrospraying has been analytically investigated. The change has been predicted by the dynamic behavior of a liquid drop at the tip of the cone-jet. Conservation laws are applied to determine the upward motion of the drop, and an instability model of electrified jets is used to determine the jet breakup. Finally, for the first time, the analysis enables prediction of the transition in terms of the Weber number and electric Bond number. The predictions are in good agreement with experimental data.

• Journal of the Korean Chemical Society
• /
• v.31 no.2
• /
• pp.153-161
• /
• 1987

The carcinogenicity of benzenoid hydrocarbons apparently depends strongly on the topological nature of the molecule. The existance of certain regions in the structure which are known propensity of benzenoid hydrocarbons to be carcinogenic. We try to identify the correlation between the number of potentially carcinogenic bay region in each of them and the quantity of Molecular Connectivity Index for 81 benzenoid hydrocarbons. Results indicate an excellent linear correlation between the number of bay region and the quantity of molecular connectivity index except the molecular containing single bond in their structure.

• 한국전산유체공학회:학술대회논문집
• /
• 2008.03b
• /
• pp.660-661
• /
• 2008

Nanotubes fabricated with diverse materials show different hydrophobic properties. The hydrophobic property is one of key properties for possible applications to ion channels due to their affinity. This study focuses on the structures of water molecules inside nanotubes with varying hydrophobicity using molecular dynamics simulation. Hydrophobicity here is determined by varying the attraction term in Lennard-Jones potential. The number of water molecules inside hydrophilic nanotubes increase, as expected, and their mobilities also increase. This trend is rather discrete with increasing number of water molecules and this discreteness is attributed to hydrogen bond. We plan to perform energy analysis to understand these structural results.

• Proceedings of the Korean Society of Propulsion Engineers Conference
• /
• 2007.04a
• /
• pp.175-178
• /
• 2007

The mode transition from cone-jet to dripping in colloid thruster operation has been analytically investigated. The transition has been predicted by the dynamic behavior of a liquid drop at the tip of the cone-jet. Conservation laws are applied to determine the upward motion of the drop, and an instability model of electrified jets is used to determine the jet breakup. Finally, for the first time, the analysis enables prediction of the transition in terms of the Weber number and electric Bond number. The predictions are in good agreement with experimental data.

• Bulletin of the Korean Chemical Society
• /
• v.25 no.12
• /
• pp.1911-1916
• /
• 2004

The conformational preference of tetrahomodioxacalix[4]arenes with three different para substituents on the phenolic ring has been investigated by using ab initio molecular orbital theory (RHF/6-31$G^{\ast}$) and density functional theory (B3LYP/6-31$G^{\ast}$). The stability order is predicted to be cone > C-1,2-alternate > partial cone > 1,3-alternate > COC-1,2-alternate. The distorted cone conformation is found to be most stable in a gas phase and the calculated results are in agreement with the reported $^1$H NMR and X-ray experimental observations. The substitution of methylene with dimethyleneoxa bridges increases the size of the annulus of the molecule, its conformational mobility, and the number of hydrogen bonding patterns. The thermodynamic stability and the conformational characteristics of tetrahomodioxacalix[4]arenes are discussed in regards of the number of phenolic hydrogen bonding patterns and the polarity of a molecule. The substituent effects on the para position of the phenolic ring are also introduced.

• Journal of the korean academy of Pediatric Dentistry
• /
• v.27 no.1
• /
• pp.45-53
• /
• 2000

Enamel microabrasion is a means by which superficial enamel discoloration is removed using hydrochloric acid and fine pumice. As enamel microabrasion alone may not be sufficient in cases of deeper discoloration, composite resin restoration is recommended in areas where there is remaining discoloration. The purpose of this study was to investigate the effects of different methods and number of applications of enamel microabrasion on the shear bond strength of composite resin to enamel. Untreated control was designated as group 1. 5-second applications of a mixture of 18% HCl and fine pumice were performed 5 and 10 times on groups 2 and 3, respectively. A commercially available mixture of 10% HCl and abrasives(PREMA) was applied using a 10 : 1 gear reduction handpiece 5 and 10 times on groups 4 and 5, respectively, with each application lasting 20 seconds. After etching with 37% phosphoric acid, composite resin was bonded. Thermocycling was performed and shear bond strength was measured. The following results were obtained : 1. Group 2 showed the highest bond strength$(24.36{\pm}3.34)$, while group 3 showed the lowest$(19.35{\pm}3.43)$, Shear bond strength decreased in the following order: 2>4>5>1>3. 2. Group 2 showed bond strength significantly higher compared to groups 1 and 3(p<0.05). 3. There were no significant differences between groups 2 and 3, which had been microabraded using HCl and pumice, and groups 4 and 5, to which PREMA had been applied, when bond strengths were compared(p>0.05). 4. When modes of fracture were examined, adhesive failure was observed in groups 3 and 4, while cohesive failure was observed in groups 1, 2, 3 and 4. Only mixed failures were found group 5. 5. When viewed using a SEM, groups 2 and 3, which had been microabraded using HCl and pumice, showed surface appearances similar to that of enamel etched with phosphoric acid. Groups 4 and 5, treated with PREMA, exhibited a smooth surface similar to that of group 1. All oops showed similar, typical surface characteristics following phosphoric acid etching.

• The korean journal of orthodontics
• /
• v.29 no.4 s.75
• /
• pp.445-456
• /
• 1999

As a result of increased education and communication, the field of orthodontics has recently been expanded to include a greater number of adult treatment procedures. With this increased demand for adult orthodontic treatment, a problem that frequently arises is the placement of appliances on teeth restored with porcelain. But conventional acid-etching is ineffective in the preparation of porcelain surface for mechanical retention of orthodontic attachments. Also, it is possible to damage on porcelain. The purpose of this study was to evaluate the effect of composite bonding materials and the porcelain surface treatment methods on shear bond strength, and to observe the porcelain fracture rates. To accomplish this purpose, this study was carried out with feldsphatic porcelain, Ceram II. Porcelain surface treatment methods were divided into intact glazed porcelain which had not treatment and surface roughening. Surface roughening by etching with Hydroluoric acid(HF), sandblasting with Microetcher II and compound treatment with etching and sandblasting. Bonding materials were Ortho-two and Transbond. All porcelain specimens were applicated with porcelain primer. 1. In comparision according to porcelain surface treatment, surface roughening groups by HF etching and sandblasting had higher shear bond than intact group. No significant difference was found in Transbond group. 2. Ortho-two group had the higher shear bond strength than that of Transbond group in B:.u etching and sandblasting. 3. E(Transbond. Intact)group had the lowest shear bond strength in all experimental group. The bond strength was higher than clinically successful bond strength. 4. Non-treated group had very higher porcelain rates than treated group. 5. This study indicates that porcelain surface-roughening may not be necessary to attachment of orthodontic brackets to porcelain surfaces.