• Communications for Statistical Applications and Methods
• /
• v.29 no.2
• /
• pp.251-261
• /
• 2022

As AI has a wide range of influence on human social life, issues of transparency and ethics of AI are emerging. In particular, it is widely known that due to the existence of historical bias in data against ethics or regulatory frameworks for fairness, trained AI models based on such biased data could also impose bias or unfairness against a certain sensitive group (e.g., non-white, women). Demographic disparities due to AI, which refer to socially unacceptable bias that an AI model favors certain groups (e.g., white, men) over other groups (e.g., black, women), have been observed frequently in many applications of AI and many studies have been done recently to develop AI algorithms which remove or alleviate such demographic disparities in trained AI models. In this paper, we consider a problem of using the information in the sensitive variable for fair prediction when using the sensitive variable as a part of input variables is prohibitive by laws or regulations to avoid unfairness. As a way of reflecting the information in the sensitive variable to prediction, we consider a two-stage procedure. First, the sensitive variable is fully included in the learning phase to have a prediction model depending on the sensitive variable, and then an imputed sensitive variable is used in the prediction phase. The aim of this paper is to evaluate this procedure by analyzing several benchmark datasets. We illustrate that using an imputed sensitive variable is helpful to improve prediction accuracies without hampering the degree of fairness much.

• Communications for Statistical Applications and Methods
• /
• v.26 no.2
• /
• pp.175-189
• /
• 2019

The problem of the estimation of quantitative sensitive variable using the item sum technique (IST) on successive occasions has been discussed. IST difference, IST regression, and IST general class of estimators have been proposed to estimate quantitative sensitive variable at the current occasion in two occasion successive sampling. The proposed new estimators have been elaborated under Trappmann et al. (Journal of Survey Statistics and Methodology, 2, 58-77, 2014) as well as Perri et al. (Biometrical Journal, 60, 155-173, 2018) allocation designs to allocate long list and short list samples of IST. The properties of all proposed estimators have been derived including optimum replacement policy. The proposed estimators have been mutually compared under the above mentioned allocation designs. The comparison has also been conducted with a direct method. Numerical applications through empirical as well as simplistic simulation has been used to show how the illustrated IST on successive occasions may venture in practical situations.

• Communications for Statistical Applications and Methods
• /
• v.8 no.3
• /
• pp.753-765
• /
• 2001

We suggest a quantitative conditional unrelated question model that can be used in obtaining more sensitive information. For whom say "yes" about the less 7han sensitive question .B we ask only about the more sensitive variable X. We extend our model to two sample case when there is no information about the true mean of the unrelated variable Y. Finally we compare the efficiency of our model with that of Greenberg et al.′s.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.25 no.12
• /
• pp.1586-1593
• /
• 1988

A new control scheme for the variable-structure control system using new time-varying switching variable is presented in this paper. It is proposed to have new algorithm for reducing the reaching time on a switching hyperplane by modifying the Morgan's algorithm. From the results of the simulation, it is concluded the proposed control algorithm yields smaller control inputs (without disturbance) and ripples (with disturbance) than that obtained by Morgan's algorithm in the steady-state. This control algorithm can be applied to proper control systems having sensitive effects on disturbances, due to the robustness.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.34 no.10
• /
• pp.389-398
• /
• 1985

Recently maximum power demand of our country has become to be under the great in fluence of electric cooling and air conditioning demand which are sensitive to weather conditions. This paper presents the technique and algorithm to forecast the long-term maximum power demand considering the characteristics of electric power and weather variable. By introducing a weather load model for forecasting long-term maximum power demand with the recent statistic data of power demand, annual maximum power demand is separated into two parts such as the base load component, affected little by weather, and the weather sensitive load component by means of multi-regression analysis method. And we derive the growth trend regression equations of above two components and their individual coefficients, the maximum power demand of each forecasting year can be forecasted with the sum of above two components. In this case we use the coincident dry bulb temperature as the weather variable at the occurence of one-day maximum power demand. As the growth trend regression equation we choose an exponential trend curve for the base load component, and real quadratic curve for the weather sensitive load component. The validity of the forecasting technique and algorithm proposed in this paper is proved by the case study for the present Korean power system.

• Communications for Statistical Applications and Methods
• /
• v.16 no.1
• /
• pp.169-183
• /
• 2009

The simplicity and wide application of Greenberg et al. (1971) prompts to propose a set of alternative estimators of population total for multi-character surveys that elicit simultaneous information on many. sensitive study variables. The proposed estimators take into account the already known rough value of the correlation coefficient between Y(the characteristic under study) and p(the measure of size). These estimators are biased, but it is expected that the extent of bias will be smaller, since the proposed estimators are suitable for situations in between those optimum for the usual estimators and the estimators based on multi-characters for no correlation. The relative efficiency of the proposed estimators has been studied under a super population model through empirical study. It has been found through simulation study that a choice of an unrelated variable in the Greenberg et al. (1971) model could be made based on its correlation with the auxiliary variable used at estimation stage in multi-character surveys.

• Journal of the Optical Society of Korea
• /
• v.8 no.2
• /
• pp.83-89
• /
• 2004

A variable optical attenuator with a bending-sensitive fiber (BSF) that can be used in optical networks is developed. The refractive index profile of the BSF is divided into four regions which are inner core, center dip of inner core, outer core and clad. The 3-dimensional finite difference beam propagation method (3D FD-BPM) is utilized to find the characteristics of the BSF, so the mode profile of the BSF and optical power attenuation according to the bending are investigated, and the equivalent model of the BSF is made. By using this equivalent model of the BSF, the BSF is fabricated, and the refractive index profile of the BSF is measured, which is similar to refractive index profile of the proposed BSF. The fabricated variable optical fiber attenuator (VOFA) consists of the BSF in a rectangular rubber ring with a fixed bend radius (BR) in a steady state. The VOFA using the proposed BSF was able to attenuate the optical power by more than about －38 ㏈ at certain wavelengths (1540∼1560 nm) based on adjusting the mechanical pressure applied to the upper surface of the rectangular rubber ring with the bent BSF. In addition, the proposed VOFA produced an insertion loss of 0.68 ㏈, polarization dependent loss (PDL) of about 0.5 ㏈, and return loss of less than －60 ㏈.

• Journal of the Korea Society for Simulation
• /
• v.22 no.3
• /
• pp.81-87
• /
• 2013

This paper is about study how to decrese surge pressure that is occurred in excavator driving motor. We used computer simulation program SimulationX. It is also about the way finding design problem and approaching a solution through interpreting shape design sensitivity analysis. Programmes are below. First of all, finding shape fault by analyzing dynamic behavior of valves installed in hydraulic driving motor which is designed now. And drawing variable which is considered sensitive to improve dynamic efficiency among a lot of shape variables. Then, targeting that variable and examining dynamic efficiency stabilization tendency with controlling it. Finally, suggesting the most effective tuning method through variable combination as there are a lot of sensitive variables.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.8 no.6
• /
• pp.1325-1331
• /
• 2007

In order to investigate the cooling capacity of an air-conditioner applying a variable capacity compressor and the temperature characteristics on a condenser and an evaporator, the experiment on the operation characteristics of the air-conditioner was performed along a compressor operation ratio and an indoor/outdoor temperatures, under a cooling operation mode. The system characteristics were measured by the psychrometric calorimeter. The cooling capacity increased with decreasing the outdoor temperature and increasing the indoor temperature. Also, it increased with increasing the compressor operation ratio. The temperature of the condenser was more sensitive for the variation of the outdoor temperature and the temperature of the evaporator was more sensitive for the variation of the indoor temperature. The operation characteristics of the cycle used in this present were also analyzed by a pressure-enthalpy chart.

• 한국정보통신설비학회:학술대회논문집
• /
• 2009.08a
• /
• pp.382-386
• /
• 2009

Online or sequential learning is one of the most basic and powerful method to train neuron network, and it has been widely used in disease detection, weather prediction and other realistic classification problem. At present, there are many algorithms in this area, such as MRAN, GAP-RBFN, OS-ELM, SVM and SMC-RBF. Among them, SMC-RBF has the best performance; it has less number of hidden neurons, and best efficiency. However, all the existing algorithms use signal normal distribution as kernel function, which means the output of the kernel function is same at the different direction. In this paper, we use multi-variable normal distribution as kernel function, and derive EKF learning formulas for multi-variable normal distribution kernel function. From the result of the experience, we can deduct that the proposed method has better efficiency performance, and not sensitive to the data sequence.