• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.38B no.12
• /
• pp.954-961
• /
• 2013

In this paper, we propose a new learning method using a variable learning to improve pattern recognition in the FCSR(Fast Commit Slow Recode) learning method of the Fuzzy ART. Traditional learning methods have used a fixed learning rate in updating weight vector(representative pattern). In the traditional method, the weight vector will be updated with a fixed learning rate regardless of the degree of similarity of the input pattern and the representative pattern in the category. In this case, the updated weight vector is greatly influenced from the input pattern where it is on the boundary of the category. Thus, in noisy environments, this method has a problem in increasing unnecessary categories and reducing pattern recognition capacity. In the proposed method, the lower similarity between the representative pattern and input pattern is, the lower input pattern contributes for updating weight vector. As a result, this results in suppressing the unnecessary category proliferation and improving pattern recognition capacity of the Fuzzy ART in noisy environments.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.12 no.11
• /
• pp.2069-2076
• /
• 2008

In this paper, we develop an electrostatic monitoring system which is composed of an electrostatic prediction system and a warning message transmission system. The electrostatic prediction system in a factory receives the value of electrostatic charge from the electrostatic sensor and predicts the next value by using past data and sends the value to the warning message transmission system through the bluetooth communication. The warning message transmission system gets a warning signal and transmits the warning message to the worker's cellphone through a commercial SMS web by a socket program running on Windows PC in a control room. We propose electrostatic forecasting algorithms based on LSR(least square regression) using weight factors in an electrostatic prediction system. Simulation results show that the algorithm with dynamically variable weight factors is best with 64.69V standard deviation and a warning message transmitted by the warning message transmission system is displayed on cellphone after about 5 seconds.

• Nutrition Research and Practice
• /
• v.5 no.4
• /
• pp.337-348
• /
• 2011

Weight reduction behavior is common among adolescent girls. The present study examined the status of weight reduction behavior and factors affecting the behavioral intention of weight reduction using the Health Belief Model (HBM) for female middle school students by weight category. Survey data was collected from three girl's middle schools in Gyeongju, Korea. A total of 299 female middle school students participated in this study. The questionnaire had questions about general characteristics, weight reduction behavior, and variables of HBM (perceived threat, perceived benefits, perceived barriers, cues to action, self-efficacy in dietary life and exercise, and behavioral intention of weight reduction). Descriptive statistics, Chi-square tests, ANOVA, and multiple regression analysis were applied to analyze the variables. A higher percentage of students in the overweight group attempted weight reduction than those in the underweight and the normal weight groups (P < 0.001). Among students who had attempted weight reduction, 73% tried diet therapy, while 78% tried exercise. Students in the normal and overweight groups showed significantly higher threat (P < 0.01) and cues to action (P < 0.001) than those in the underweight group. As for perceived benefits, barriers, and self-efficacy in dietary life and exercise, there were no significant differences among weight groups. Students in the overweight group showed the strongest intention of weight reduction and there were significant differences among the three weight groups (P < 0.001). Perceive threat (P < 0.01), cues to action (P < 0.001), and perceived self-efficacy (P < 0.01) were significantly associated to behavioral intention of weight reduction for all respondents. For the underweight group, perceived threat (P < 0.05) and perceived self-efficacy (P < 0.01) were the significant variables. For the overweight group, cue to action was the significant variable (P < 0.05).

• Proceedings of the Korea Air Pollution Research Association Conference
• /
• 2005.11a
• /
• pp.100-102
• /
• 2005

• Communications of the Korean Mathematical Society
• /
• v.35 no.2
• /
• pp.517-532
• /
• 2020

In this paper, we consider a class of nonlocal problems with indefinite weights in Orlicz-Sobolev space. Under some suitable conditions on the nonlinearities, we establish some existence results using variational techniques and Ekeland's variational principle.

• Asian-Australasian Journal of Animal Sciences
• /
• v.3 no.4
• /
• pp.263-267
• /
• 1990

The fat thicknesses of twenty-eight Japanese Black beef steers were measured with an ultrasonic device at eleven points on the cattle prior to slaughter and side dissection. The relation between live fat thickness and both weight and percentage of fat and lean in the carcass was examined. Fat thickness obtained from nine points of the chest, flank and rump regions was found to relate significantly (P<0.01) to both weight and percentage of fat. However, shoulder fat thickness measurements were not significantly related to the weight or percentage of fat or lean in the carcass. Addition of live fat thickness to animal age or live weight as an independent variable markedly improved the precision of multiple regression equations for predicting weight of fat and lean, and percentage of fat. In predicting the percentage of lean, both animal age and body weight were not employed in the multiple regression equation. The residual standard deviation for predicting percentage of fat and lean were 1.93 and 1.87, respectively. The ultrasonic measurement of fat thickness if supposed to be useful to the prediction of carcass composition of beef cattle.

• Journal of Life Science
• /
• v.7 no.4
• /
• pp.262-269
• /
• 1997

There are many different views of the opinion that smoking decreased the HDL cholesterol. So this cross-sectional study was attempted to find the influence which smoking, alcohol and weight affect HDL cholesterol. They were two groups;smoking group(409 males) and nonsmoking group(213 males). And HDL cholesterol, triglyceride, and total cholesterol of smoking group were compaired to them of nonsmoking group. Also the variables of age, relative weight, and alcohol were compaired. The HDL cholesterol of nonsmoking group was 47.2$\pm$10.1 mg/dl and the HDL cholesterol of smoking group was 45.5$\pm$10.7mg/dl. So there were no statistical difference. As smoking was increased in amount, HDL cholesterol was decreased slightly. According as the alcohol intake increases, especially only the HDL cholesterol of the moderate alcohol intake(101-200gm/week) group was significantly increased(P<0.05). As relative weight increases, HDL cholesterol were significantly decreased(P=0.007). Smoking was not statistical difference of HDL cholesterol. Moderately alcohol intake group(101-200gm/week) were increased the HDL cholesterol(P<0.05). Only relative weight was statistically decreased the HDL cholesterol(P<0.01). Multiple regression results that relative weight and alcohol were significant variable of the level of HDL cholesterol.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.28 no.10
• /
• pp.1574-1582
• /
• 2004

Binary manipulators have recently received much attention due to hyper-redundancy, light weight, good controllability and high reliability. The precise positioning of the manipulator end-effecter requires the use of many modules, which results in a high-dimensional workspace. When the workspace dimension is large, existing inverse kinematics methods such as the Ebert-Uphoff algorithm may require impractically large memory size in determining the binary positions of all actuators. To overcome this limitation, we propose a new inverse kinematics algorithm: the inverse kinematics problem is formulated as an optimization problem using real-valued design variables, The key procedure in this approach is to transform the integer-variable optimization problem to a real-variable optimization problem and to push the real-valued design variables as closely as possible to the permissible binary values. Since the actual optimization is performed in real-valued design variables, the design sensitivity becomes readily available, and the optimization method becomes extremely efficient. Because the proposed formulation is quite general, other design considerations such as operation power minimization can be easily considered.

• Journal of KSNVE
• /
• v.9 no.1
• /
• pp.121-130
• /
• 1999

Generally, A Dynamic Absorber by using Active viscous Damping is highlighted for effective suspension system, such as improved ride comfort and handling in the market. Lately, this system based on the Sky-Hook damper theory is introduced by the name of "Active Dynamic Absorber" to us. This system has an excellent performance in contrast to Passive. Adaptive Dynamic Absorber, besides having low cost components of system, low energy consumption. light weight of system. In this viewpoint. most of car-maker will adopt this system in the near future. For this reason, we developed Dynamic Absorber by using Active viscous Damping which is equipped with continuously variable Dynamic Absorber and Control logic consisting Filter and Estimator. control apparatus of Dynamic Absorber operated by 16-bit microprocessor of high performance. variable device of viscous Damping. G-sensor so on. In this paper. several important points of development procedure for realizing this system will be described with results in which is obtained from experiment by simulation and Full car test in Proving ground. respectively.pectively.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.37 no.10
• /
• pp.1195-1200
• /
• 2013

In recent years, variable stiffness actuation has attracted much attention because interaction between a robot and the environment is increasingly required for various robot tasks. Several variable stiffness actuators (VSAs) have been developed; however, they find limited applications owing to their size and weight. For realizing their widespread use, we developed a compact and lightweight mini-VSA. The mini-VSA consists of a control module based on an adjustable moment arm mechanism and a drive module with two motors. By controlling the relative motion of cams in the control module, the position and stiffness can be simultaneously controlled. Experimental results are presented to show its ability to change stiffness.