• The Korean Journal of Applied Statistics
• /
• v.34 no.5
• /
• pp.711-721
• /
• 2021

In this paper, we propose the method of adjusting thresholds using impurity indices in classification analysis on imbalanced data. Suppose the minority category is Positive and the majority category is Negative for the imbalanced binomial data. When categories are determined based on the commonly used 0.5 basis, the specificity tends to be high in unbalanced data while the sensitivity is relatively low. Increasing sensitivity is important when proper classification of objects in minority categories is relatively important. We explore how to increase sensitivity through adjusting thresholds. Existing studies have adjusted thresholds based on measures such as G-Mean and F1-score, but in this paper, we propose a method to select optimal thresholds using the chi-square statistic of CHAID, the Gini index of CART, and the entropy of C4.5. We also introduce how to get a possible unique value when multiple optimal thresholds are obtained. Empirical analysis shows what improvements have been made compared to the results based on 0.5 through classification performance metrics.

• Journal of the Korean Society for Nondestructive Testing
• /
• v.25 no.5
• /
• pp.339-342
• /
• 2005

In long-period fiber grating (LPFG) to be made up optical fiber sensors, resonance coupling occurs between the forward-propagating core mode and cladding modes at the wavelength that satisfy the Phase matching condition. The resonance wavelength and the coupling strength depends strongly on the external environment like temperature, strain, and ambient index. These characteristics can be utilized for various applications as optical fiber sensors. fabrication of optical fiber gratings is typically based on the photosensitivity effect, i.e. the permanent change of the refractive index upon irradiation of the UV beam, and therefore, fabrication of the optical fiber with high phososensitivity is an important part of the research on optical fiber gratings. In this work, we measured the effort of to-doping of boron on the index difference between the core and cladding of the optical fiber and the sensitivity of the LPFC to the temperature and bending changes. We observed that the index difference between the core and the cladding decreased by $(1.69{\times}10^{-4}/SCCM)$ and the temperature sensitivity of the resonance wavelength shirt decreased by $(0.01145nm/^{\circ}C/SCCM)$. The dependence or the bending-induced changes or the transmission characteristics of LPFG on the tore-cladding index difference was investigated experimentally. The measurement results indicate that the bending sensitivity increases as the index difference decreases.

• Current Optics and Photonics
• /
• v.4 no.1
• /
• pp.44-49
• /
• 2020

In this paper, a new method is presented to measure the refractive index of single plain glass or multilayered materials, based on a laser frequency-shifted confocal feedback microscope. Combining the laser frequency-shifted feedback technique and the confocal effect, the method can attain high axial-positioning accuracy, stability and sensitivity. Measurements of different samples are given, including N-BK7 glass, Silica plain glass, and a microfluidic chip with four layers. The results for N-BK7 glass and Silica plain glass show that the measurement uncertainty in the refractive index is better than 0.001. Meanwhile, the feasibility of this method for multilayered materials is tested. Compared to conventional methods, this system is more compact and has less difficulty in sample processing, and thus is promising for applications in the area of refractive-index measurement.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.27 no.8
• /
• pp.1426-1435
• /
• 2003

Design goal is to find the one that has the highest probability of success and the smallest variation. A robustness index has been proposed to satisfy these conditions. The two-step optimization process of the target problem requires a scaling factor. The search process of a scaling factor is replaced with the making of the decoupled design between the mean and the standard deviation. The decoupled design matrix is formed from the sensitivity or the sum of squares. After establishing the design matrix, the robust design process has a new three-step one. The first is ″reduce variability,″ the second is ″make the candidate designs that satisfy constraints and move the mean on the target,″ and the final is ″select the best robust design using the proposed robustness index.″ The robust design process is verified by three examples and the results using the robustness index are compared with those of other indices.

• Textile Coloration and Finishing
• /
• v.16 no.2
• /
• pp.47-51
• /
• 2004

A differential refractive index detector was used for monitoring glucose contents in textile processing solutions. The sensitivity of the device was high enough to measure 0.05％ aqueous glucose solution that could not be measured by normal refractive index measurement. The device was set to monitor glucose concentration continuously in real time by measuring differential refractive index and calibrated by standard glucose solutions in a range of 0.1 to 1.0%. The possibility of industrial application of the device was demonstrated by real-time monitoring of glucose concentration in textile processing solutions such as desizing bath and cellulase treatment bath. Both of solutions contained glucose as a major degraded product. The device would be able to control the weight loss of cellulosic fiber during cellulase treatment since the amount of degraded products in a processing bath is proportional to its weight loss.

• ETRI Journal
• /
• v.27 no.4
• /
• pp.377-384
• /
• 2005

We investigate near-infrared light (NIR) propagation in a model of an adult head using an extensive Monte Carlo (MC) simulation. The adult head model is a four-layered slab which consists of a surface layer, a cerebrospinal fluid layer, a gray-matter layer, and a white-matter layer. We study the effects of a refractive index mismatch on the model, calculating the intensity of detected light, mean flight time, and partial mean flight time of each layer for various refractive indices of the cerebrospinal fluid layer as functions of source-detector spacing. The Monte Carlo simulation shows that the refractive index mismatch presents very rich results including rapidly decaying intensity of detected light and a peak and cross-over in the partial mean flight time with source-detector spacing. We also investigate spatial sensitivity profiles at various source-detector spacings, discussing the index mismatch effect on the model.

• Journal of Physiology & Pathology in Korean Medicine
• /
• v.23 no.5
• /
• pp.950-957
• /
• 2009

The purpose of this study was to develop generalized validity evaluation indexes for the Sasang type diagnostic test. As a generalized evaluation index for Sasang type diagnostic test, Generalized Squared Correlation (GC2), Mutual Information (I) and Mutual Information Coefficient (IC) as overall validity index, and Correlation Coefficient (C) and Mutual Information Contribution (ID) as type-specific validity index were suggested and the practicability of these indexes was assessed along with Percentage Correctly Predicted (PCP), adjusted PCP, type-specific sensitivity and type-specific specificity which was proposed previously. The usefulness of the nine validity indexes was examined using previously reported data and three mock Sasang type diagnosis results. We demonstrated here that the PCP is not adequate for the analysis of validity. We also showed that our validity index is recommendable for the comparison and improvement of Sasang type diagnosis. The generalized validity analysis of Sasang type diagnosis can provide a guideline for the development of new Sasang type diagnosis or pattern identification. The working sheet for calculating validity indexes can be acquired at http://www.sasangtypology.org/validity.

• Journal of the Korean Society of Manufacturing Technology Engineers
• /
• v.19 no.4
• /
• pp.529-538
• /
• 2010

This paper presents a reliability-based topology optimization (RBTO) based on bidirectional evolutionary structural optimization (BESO). In design of a structure, uncertain conditions such as material property, operational load and dimensional variation should be considered. Deterministic topology optimization (DTO) is performed without considering the uncertainties related to the design variables. However, the RBTO can consider the uncertainty variables because it can deal with the probabilistic constraints. The reliability index approach (RIA) and the performance measure approach (PMA) are adopted to evaluate the probabilistic constraints in this study. In order to apply the BESO to the RBTO, sensitivity number for each element is defined as the change in the reliability index of the structure due to removal of each element. Smoothing scheme is also used to eliminate checkerboard patterns in topology optimization. The limit state indicates the margin of safety between the resistance (constraints) and the load of structures. The limit State function expresses to evaluate reliability index from finite element analysis. Numerical examples are presented to compare each optimal topology obtained from RBTO and DTO each other. It is verified that the RBTO based on BESO can be effectively performed from the results.

• Sleep Medicine and Psychophysiology
• /
• v.13 no.2
• /
• pp.59-66
• /
• 2006

The objective of this study was to investigate the relationship between insomnia severity, depression, anxiety, and anxiety sensitivity and to find out the explanatory variables that account for the insomnia severity among depression, anxiety, and anxiety sensitivity in general population. 95 mentally healthy volunteers who visit health promotion center of Kangbuk Samsung hospital for their regular medical examination were requested to complete Athens Insomnia Scale, Beck Depression Inventory, State-Trait Anxiety Inventory, and Anxiety Sensitivity Index. Association between total scores of Athens Insomnia Scale and other variables (total scores of Beck Depression Inventory excluded item 16, total scores of State Anxiety, total scores of Trait Anxiety, and total scores of Anxiety Sensitivity Scale) was assessed individually with partial correlations adjusted by age and then together using multiple regression analysis. The total scores of Athens Insomnia Scale were significantly associated with total scores of Beck Depression Inventory excluded item 16 (r=0.541, p<0.001), total scores of Trait Anxiety (r=0.642, p<0.001), total scores of State Anxiety (r=0.267, p<0.05), and total scores of Anxiety Sensitivity Index (r=0.312, p<0.01). Total scores of trait anxiety showed the highest correlation with the total scores of Athens Insomnia Scale and was the significant predictor to total scores of Athens Insomnia Scale among the other predictor variables (p<0.001). These results show that insomnia severity is positively correlated with depression, anxiety, and anxiety sensitivity. The correlation was strongest with trait anxiety. In addition, our results suggest that trait anxiety is associated with insomnia severity in general populations.

• Water for future
• /
• v.27 no.1
• /
• pp.89-99
• /
• 1994

Flood forecasting in Korea has been based on the off-line parameter estimation method. But recent flood forecasting studies explore on-line recursive parameter estimation algorithms. In this study, a simultaneous adaptive estimation of system states and parameters for rainfall-runoff model is investigated for on-line real-time flood forecasting and parameter estimation. The proposed flood routing system is composed of Flood forecasting in Korea has been based on the off-line parameter estimation method. But recent flood forecasting studies explore on-line recursive parameter estimation algorithms. In this study, a simultaneous adaptive estimation of system states and parameters for rainfall-runoff model is investigated for on-line real-time flood forecasting and parameter estimation. The proposed flood routing system is composed of ø-index in the assessment of effective rainfall and the cascade of nonlinear reservoirs accounting for translation effect in flood routing. To combine the flood routing model with a parameter estimation model, system states and parameters are treated with the extended state-space formulation. Generalized least squares and maximum a posterior estimation algorithms are comparatively examined as estimation techniques for the state-space model. The sensitivity analysis is to investigate the identifiability of the parameters. The index of sensitivity used in this study is the covariance matrix of the estimated parameters.-index in the assessment of effective rainfall and the cascade of nonlinear reservoirs accounting for translation effect in flood routing. To combine the flood routing model with a parameter estimation model, system states and parameters are treated with the extended state-space formulation. Generalized least squares and maximum a posterior estimation algorithms are comparatively examined as estimation techniques for the state-space model. The sensitivity analysis is to investigate the identifiability of the parameters. The index of sensitivity used in this study is the covariance matrix of the estimated parameters.