• 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.

• Journal of the Korean Applied Science and Technology
• /
• v.37 no.5
• /
• pp.1190-1199
• /
• 2020

The purpose of this study is to analyze the variables affecting the effective stroke in the swing performed through 12 weeks of training for golf beginners, and to provide basic data on the effective stroke factors for the golf beginners to settle on the fairway. Twenty subjects were participate in the study (age: 21.35±1.69 yrs, height: 176.75±7.99 cm, weight: 70.70±9.76 kg). All subjects were subjected to a 12-week golf training, and trackman 4 was used in the 12th week to calculate variables affecting the effective stroke during a golf swing. Trackman data was divided into club-variables and ball-variables, and a binary logistic regression analysis was performed to find out the variables affecting effective strokes. In club-variables, high dynamic loft and low face angles were found in effective stroke, and in ball-variables, fast ball speed, large smash factor, high launch angle, and many spin rates were also found in effective stroke. As a result of the binary logistic regression analysis of club-variables, the probability of an effective stroke increased as the club speed and dynamic loft increased, and the probability of an effective stroke decreased as the face angle increased. The influence of effective stroke in the club-variables was in the order of dynamic loft, face angle, and club speed. In the ball-variable, the probability of an effective stroke increased when the lunch angle increased, and the probability of an effective stroke decreased as the lunch direction increased. As a condition to increase the probability of effective stroke based on the results, it is necessary to increase the club speed through high dynamic loft and low face angle during swing through continuous practice. Through this, the probability of effective stroke through increasing the launch angle and decreasing the launch direction will increase.

• Journal of the Korean Society of Food Science and Nutrition
• /
• v.5 no.1
• /
• pp.87-92
• /
• 1976

Comparing the amount of Hb, the nutriture of the deaf-mutes are better than that of the amentias and the nursery school children. The former, however, shows less condition than that of the boys and girls in the junior physical education school (Hb : 12.7) and the normal family children (Hb : 10.8). In the group of six to eleven years old that is a class a elementary school, the physical condition such as stature, chest circumference, and weight of the deaf -mute is respectively almost the same degree, which is above standard in our country. On the contrary the degree of the amentias and the nursery school children has an inferiority approaching to the standard. In the group of twelve to fifteen years old that is a class of junior high school, the nursery school children have the lowest physical condition and the boys and girls of the physical training school the highest. The order of the growth of physical condition is as follows : Nursery school children

• Korean Journal of Environmental Agriculture
• /
• v.24 no.1
• /
• pp.66-70
• /
• 2005

The experiment was carried out to investigate the effects of gamma radiation on the early growth performance and physiological traits of BT2 clone tea, the most promising cultivar released by Bangladesh Tea Research Institute. The fresh shoot cuttings were irradiated with seven different levels of gamma radiation such as 0, 10, 20, 30, 40, 50 and 60 Gy from Cobalt 60Co source (Dept. of PlantBreeding, Bangladesh Institute of Nuclear Agriculture). Thereafter, the irradiated shoot cuttings were planted in polythene bags and kept under natural conditions. It was observed that callusing was initiated from 8th weeks after placement of tea shoot cuttings in the polythene bags and completed by 12th weeks. The morphological growth of tea shoot cuttings were recorded under varying levels of gamma radiation and growth stages. It was observed that the number of leaves, number of primary branches, base diameter, root length and total leaf area per plant significantly increased with the progress of time and increasing levels of gamma radiation, however, the plant height showed decreasing trend with the increasing levels of gamma radiation, which could be due to the change in chromosomal structure and genetic makeup. After 56 weeks of planting, the plant height, the number of leaves and primary branches per plant, base diameter, root length and total leaf area per plant recorded were 65.70 cm, 30.67, 7.33, 1.48 cm, 23.50 cm, and 1250.67 cm2 per plant respectively under the radiation level 60 Gy, whereas the corresponding figures of the above parameters at the control treatment were 76.21 cm, 18.33, 3.67, 0.92 cm, 17.75 cm and 778.33 cm2 per plant, respectively. A significant relationship was observed among the physiological growth parameters with the increasing levels of gamma radiation. The total dry matter gain, leaf area index, absolute growth rate and relative growth rate were significantly influenced with the rising levels of gamma radiation (up to 60 Gy), whereas the net assimilation rate of individual tea plant non-significantly responded as compared to those of control treatment. Finally after 56 weeks of planting, the maximum total dry weight gain, leaf area index, absolute growth rate, relative growth rate and net assimilation rate recorded under 60 Gay radiation level were 40.25 g/plant/week, 4.25, 1.18 g/week, 0.0621g/g/week and 17.07 g/m2/week respectively.

• The Transactions of the Korea Information Processing Society
• /
• v.1 no.2
• /
• pp.225-236
• /
• 1994

Many research result conducted by neural network researchers claimed that the degree of generalization of the neural network system is higher or at least equal to that of statistical methods. However, those successful results could be brought only if the neural network was trained by appropriately sound data, having a little of noisy data and being large enough to control noisy data. Real data used in a lot of fields, especially business fields, were not so sound that the network have frequently failed to obtain satisfactory prediction accuracy, the degree of generalization. Enhancing the degree of generalization with noisy data is discussed in this study. The suggestion, which was obtained through a series of experiments, to enhance the degree of generalization is to remove inconsistent data by checking overlapping and inconsistencies. Furthermore, the previous conclusion by other reports is also confirmed that the learning mechanism of neural network takes average value of two inconsistent data included in training set［2］. The interim results of on-going research project are reported in this paper These are ann architecture of the neural network adopted in this project and the whole idea of developing on-line credit evaluation system,being intergration of the expert(resoning)system and the neural network(learning system.Another definite result is corroborated through this study that quickprop,being agopted as a learing algorithm, also has more speedy learning process than does back propagation even in very noisy environment.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.19 no.4
• /
• pp.375-381
• /
• 2013

To assess the risk of marine traffic environments, with high confidential level, the risk factors comprising it should be identified and the risk acceptance criteria should be also provided. Furthermore, the relative importance of each risk factor(the weight of each risk factor on total risk) should be analyzed because the risk is expressed as the sum of risk factors comprising it. The twenty kinds of risk factors and its assessment criteria were suggested for the domestic marine traffic environments by an examination of the existing risk assessment models on the previous studies. The relative importance of each risk factor was also analyzed through the questionnaire using analytic hierarchy process by the marine traffic experts on the same studies. Based on these previous studies, the risk was evaluated at the port of Mokpo and its approaches on this study. The port of Mokpo and its approaches were divided into four sectors for the comparative evaluation, the result of the comparative evaluation on four sectors showed that the risk of the Jeongdeung-hae passage is the highest due to higher risk level of some risk factors(water movements, complexities, tug boats, pilotage, VTS) than the other sectors. The result of this evaluation is in accord with the analysis results of the other studies using various qualitative or quantitative risk analysis methods at the same sea areas.

• Journal of the Korean Data and Information Science Society
• /
• v.23 no.5
• /
• pp.895-904
• /
• 2012

This study aims to examine the effects of a 12-week compound exercise program (aerobic exercise+weight training) on obese middle school girls' leptin and insulin before and after the exercise. This is achieved by dividing obese middle school girls whose body fat percentage is over 30% into a compound exercise group (n=20) and a control group (n=20) and conducting comparative analysis on them.After the Shapiro-Wilk normality test of the variables, a two-sample t-test was performed to see if the variables have the same mean between the compound exercise and control groups. A paired t-test was also performed to see if the changes in the variables before and after the compound exercise program were statistically significant. For all the statistical analysis, the significance level was set at ${\alpha}=0.05$. The results of this study showed the leptin and insulin levels in the combined exercise group had been significantly decreased. The regular 12 weeks of combined exercise is considered to have a positive impact on leptin and insulin levels in obese schoolgirls.

• Fire Science and Engineering
• /
• v.33 no.6
• /
• pp.35-42
• /
• 2019

The aim of this study is to investigate the direction of improvement of safety experience education through the analysis of the evacuation time experiment. For the study, test subjects were divided into groups of similar body size and weight. The test subjects were directly experienced four evacuation devices, and the experience time was measured. As a result of the analysis of the total time from the installation of the evacuation device to the escape, the time was measured in the order of Descending Life Line-Tilt-Down Rescue Line-Vertical Escape Chute-air safety mat. In the case of evaluating the evacuation time using evacuation mechanisms, the evacuation time was measured in the order of air safety mat-Tilt-Tilt-Down Rescue Line-Descending Life Line-Vertical Escape Chute. In the first and second experiments of the Descending Life Line, time differences were observed. The escape time using the Descending Life Line was reduced in the second experiment than in the first experiment. As shown in this result, education through experience has shown that behavioral confidence and time can be managed. The conclusion of this study is that the goal of safety education is to minimize human life and property damage. Therefore, in order to bring this effect to more people, it is necessary to make efforts to keep self-safe through experiential education.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.24 no.1
• /
• pp.79-85
• /
• 2011

Bridge Weigh-in-Motion(BWIM) system calculates a travelling vehicle's weight without interruption of traffic flow by analyzing the signals that are acquired from various sensors installed in the bridge. BWIM system or data accumulated from the BWIM system can be utilized to development of updated live load model for highway bridge design, fatigue load model for estimation of remaining life of bridges, etc. Field test with moving trucks including various load cases should be performed to guarantee successful development of precise BWIM system. In this paper, a numerical simulation technique is adopted as an alternative or supplement to the vehicle traveling test that is indispensible but expensive in time and budget. The constructed numerical model is validated by comparison experimentally measured signal with numerically generated signal. Also vehicles with various dynamic characteristics and travelling conditions are considered in numerical simulation to investigate the variation of bridge responses. Considered parameters in the numerical study are vehicle velocity, natural frequency of the vehicle, height of entry bump, and lateral position of the vehicle. By analyzing the results, it is revealed that the lateral position and natural frequency of the vehicle should be considered to increase precision of developing BWIM system. Since generation of vehicle travelling signal by the numerical simulation technique costs much less than field test, a large number of test parameters can effectively be considered to validate the developed BWIM algorithm. Also, when artificial neural network technique is applied, voluminous data set required for training and testing of the neural network can be prepared by numerical generation. Consequently, proposed numerical simulation technique may contribute to improve precision and performance of BWIM systems.

• Physical Therapy Korea
• /
• v.25 no.2
• /
• pp.22-29
• /
• 2018

Background: Deficits of both ankle dorsiflexion range of motion (DFROM) and dynamic balance are shown in persons with chronic ankle instability and the elderly, with the risk of falls. Objects: This study aims to investigate the relationship between DFROM and dynamic balance in elderly subjects and young adults. Methods: Fifty-nine subjects were divided into three groups: ankle stability young group (SY), ankle instability young group (IY) and ankle stability older group (SO). We recruited three old subjects with ankle instability, but excluded them during a pilot testing due to the safety issue. DFROM was measured by weight bearing lunge test (WBLT) and dynamic balance was measured via star excursion balance test (SEBT) in anteromedial, medial, and posteromedial directions. The group differences in WBLT and SEBT and each group's correlation between WBLT and SEBT were detected using the R statistical software package. Results: The dorsiflexion range of motion was significantly different between the SY, IY, and SO groups. The SO group showed the highest DFROM and IY group showed the lowest DFROM (SY: $45.88{\pm}.66^{\circ}$, IY: $39.53{\pm}1.63^{\circ}$, SO: $47.94{\pm}.50^{\circ}$; p<.001). However, the SO group showed the lowest dynamic balance score for all SEBT directions (SY: $87.24{\pm}2.05cm$, IY: $83.20{\pm}1.30cm$, SO: $77.23{\pm}2.07cm$; p<.05) and there was no relationship between the dorsiflexion range of motion and dynamic balance in any group. Conclusion: Our findings suggest that ankle DFROM is not a crucial factor for dynamic stability regardless of aging and ankle instability. Other factors such as muscle strength or movement coordination should be considered for training dynamic balance. Therefore, we need to establish the rehabilitation process by measuring and treating ROM, balance, and muscle strength when treating young adults with and without ankle instability as well as elderly people.