• Communications for Statistical Applications and Methods
• /
• v.9 no.2
• /
• pp.411-423
• /
• 2002

In this paper, we propose a split criterion for binary decision trees. The proposed criterion selects the optimal split by measuring the prediction success of the candidate splits at a given node. The criterion is shown to have the property of exclusive preference. Examples are given to demonstrate the properties of the criterion.

• Communications for Statistical Applications and Methods
• /
• v.10 no.2
• /
• pp.471-478
• /
• 2003

In this paper, we reconsider the Kolmogorov-Smirnoff distance as a split criterion for binary decision trees and suggest an algorithm to obtain the Kolmogorov-Smirnoff distance more efficiently when the input variable have more than three categories. The Kolmogorov-Smirnoff distance is shown to have the property of exclusive preference. Empirical results, comparing the Kolmogorov-Smirnoff distance to the Gini index, show that the Kolmogorov-Smirnoff distance grows more accurate trees in terms of misclassification rate.

• Advanced Composite Materials
• /
• v.16 no.2
• /
• pp.167-180
• /
• 2007

The tensile strength of unidirectional carbon fiber reinforced plastics under a high strain rate was experimentally investigated. A high-strain-rate test was performed using the tension-type split Hopkinson bar technique. In order to obtain the tensile stress-strain relations, a special fixture was used for the impact tensile specimen. The experimental results demonstrated that the tensile modulus and strength in the longitudinal direction are independent of the strain rate. In contrast, the tensile properties in the transverse direction and the shear properties increase with the strain rate. Moreover, it was observed that the strain-rate dependence of the shear strength is much stronger than that of the transverse strength. The tensile strength of off-axis specimens was measured using an oblique tab, and the experimental results were compared with the tensile strength predicted based on the Tsai-Hill failure criterion. It was concluded that the tensile strength can be characterized quite well using the above failure criterion under dynamic loading conditions.

• Advanced Composite Materials
• /
• v.17 no.2
• /
• pp.139-156
• /
• 2008

This study aims to characterize the dynamic tensile strength of unidirectional carbon/epoxy composites. Two different carbon/epoxy composite systems, the unidirectional T700S/2500 and TR50S/modified epoxy, are tested at the static condition and the strain rate of $100\;s^{-1}$. A high-strain-rate test was performed using a tension-type split Hopkinson bar technique with a specific fixture for specimen. The experimental results demonstrated that both tensile strength increase with strain rate, while the fracture behaviors are quite different. By the use of the rosette analysis and the strain transformation equations, the strain rate effects of material principal directions on tensile strength are investigated. It is experimentally found that the shear strain rate produces the more significant contribution to strain rate effect on dynamic tensile strength. An empirical failure criterion for characterizing the dynamic tensile strength was proposed based on the Hash-in's failure criterion. Although the proposed criterion is just the empirical formula, it is in better agreement with the experimental data and quite simple.

• Computers and Concrete
• /
• v.22 no.3
• /
• pp.291-303
• /
• 2018

As the concept of "sponge city" is proposed, the pervious concrete for permeable pavement has been widely used in pavement construction. This paper aims at investigating the dynamic behavior and energy evolution of pervious concrete under impact loading. The dynamic compression and split tests are performed on pervious concrete by using split Hopkinson pressure bar equipment. The failure criterion on the basis of incubation time concept is used to analyze the dynamic failure. It is demonstrated that the pervious concrete is of a strain rate sensitive material. Under high strain rate loading, the dynamic strength increases while the time to failure approximately decreases linearly as the strain rate increases. The predicted dynamic compressive and split tensile strengths based on the failure criterion are in accordance with the experimental results. The total damage energy is found to increase with the increasing of strain rate, which means that more energy is needed to produce irreversible damage as loading rate increases. The fractal dimensions are observed increases with the increasing of impact loading rate.

• Journal of Industrial Technology
• /
• v.23 no.A
• /
• pp.139-147
• /
• 2003

The purpose of this study was to investigate and analyze the fatigue test data of pavement concrete. The static strength tests were carried out to check the compressive strength, flexural strength, and split tension strength at 56 days in order to minimize strength variation effect during test. The specimens were fabricated at twelves sections at a construction site of highway. The stress level and stress ratio of fatigue test were determined from static test results. The results are as follow: The flexural strength at 28 days mostly satisfied the criterion for design, but the compressive strength at 28 days were slightly below the criterion even though it satisfied at 56 days. The fatigue limit was 2 million cycles if the specimen was not failed to that cycles. The S-N curves were developed from the fatigue test results at each stress levels and each stress ratio. Then, the fatigue life of pavement concrete at a given stress level and fatigue strength of pavement concrete could be derived from these curves. Analysis using method No.2 was more acceptable because resulting of comparison and analysis using method No.2 was presented 2 sections were presented $R^2$ < 0.7, and other 2 sections were presented 0.7 < $R^2$ < 0.8, and the others 8 sections were $R^2{\geq}0.8$.

• Communications for Statistical Applications and Methods
• /
• v.10 no.3
• /
• pp.627-635
• /
• 2003

In this short communication we discuss the bias problem of C4.5 in split variable selection and suggest a method to reduce the variable selection bias among categorical predictor variables. A penalty proportional to the number of categories is applied to the splitting criterion gain of C4.5. The results of empirical comparisons show that the proposed modification of C4.5 reduces the size of classification trees.

• The Journal of the Acoustical Society of Korea
• /
• v.28 no.3E
• /
• pp.118-135
• /
• 2009

The level set based approach is one of active methods for contour extraction in image segmentation. Since Osher and Sethian introduced the level set framework in 1988, the method has made the great impact on image segmentation. However, there are some problems to be solved; such as multi-objects segmentation, noise filtering and much calculation amount. In this paper we address the drawbacks of the previous level set methods and propose an extension of the traditional fast level set to cope with the limitations. We introduce a relationship matrix, a new split-and-merge criterion, a modified Chan-Vese criterion and a novel filtering criterion into the traditional fast level set approach. With the segmentation experiments we evaluate the proposed method and show the promising results of the proposed method.

• The Korean Journal of Applied Statistics
• /
• v.24 no.5
• /
• pp.759-768
• /
• 2011

Tree algorithms have been widely developed for regression problems. One of the good features of a regression tree is the flexibility of fitting because it can correctly capture the nonlinearity of data well. Especially, data with sudden structural breaks such as the price of oil and exchange rates could be fitted well with a simple mixture of a few piecewise linear regression models. Now that split points are determined by chi-squared statistics related with residuals from fitting piecewise linear models and the split variable is chosen by an objective criterion, we can get a quite reasonable fitting result which goes in line with the visual interpretation of data. The piecewise linear regression by a regression tree can be used as a good fitting method, and can be applied to a dataset with much fluctuation.

• Journal of Korean Academy of Nursing
• /
• v.46 no.3
• /
• pp.431-442
• /
• 2016

Purpose: This paper was conducted to validate the Korean version of the Peer Group Caring Interaction Scale (PGCIS-K) that measures caring behaviors as experienced by nursing students. Methods: Translation of the PGCIS-K was validated through forward-backward translation methods. Survey data were collected from 218 nursing students in a nursing school. Construct validity and criterion-related validity were evaluated. Internal consistency and the Guttman split-half coefficient were calculated to assess reliability. Results: The PGIS-K showed reliability except for 4 items (Cronbach's ${\alpha}=.91$, Guttman split-half coefficient=.85), which were low (<.30) or negatively correlated with the total scale. A 12-item reduced form of the PGCIS-K was developed by item-analysis and construct validity evidence. Factor loading for the 12 items on 2 factors ranged from .47~.82, which explained 58.4% of the total variance. Two factors were named 'modeling and assistance (Cronbach's ${\alpha}=.87$)' and 'communication and sharing (Cronbach's ${\alpha}=.82$)'. Convergent validity, discriminant validity, and criterion validity were supported according to the correlation coefficients of the 2 factors with other measure. Conclusion: The findings suggest preliminary evidence that the 12-item PGCIS-K can be used to measure nursing students' peer group caring interactions in Korea. Additional studies are recommended to continue the psychometric evaluation of this scale. Also, it can be extended to measure graduate nursing students or staff nurses' peer group caring interaction.