• 한국경영과학회지
• /
• 제39권4호
• /
• pp.75-84
• /
• 2014

In this paper, we dealt with feature selection problem of large-scale and high-dimensional biological data such as omics data. For this problem, most of the previous approaches used simple score function to reduce the number of original variables and selected features from the small number of remained variables. In the case of methods that do not rely on filtering techniques, they do not consider the interactions between the variables, or generate approximate solutions to the simplified problem. Unlike them, by combining set covering and clustering techniques, we developed a new method that could deal with total number of variables and consider the combinatorial effects of variables for selecting good features. To demonstrate the efficacy and effectiveness of the method, we downloaded gene expression datasets from TCGA (The Cancer Genome Atlas) and compared our method with other algorithms including WEKA embeded feature selection algorithms. In the experimental results, we showed that our method could select high quality features for constructing more accurate classifiers than other feature selection algorithms.

• 산업경영시스템학회지
• /
• 제29권3호
• /
• pp.7-14
• /
• 2006

Nowadays, many people are deeply interested in health and well-being. Ministry of Health and Welfare showed the results of evaluation of medical service of seventy eight general hospitals in 2005. The number of its medical service variables was eighteen and the evaluation was carried out in eight regions on a national scale. Our study used the result data of the evaluation and presented the descriptive and inferential research findings. Thus, relations between the eight regions in Korea and 18 variables were studied, using ANOVA, etc., the significant results were found, and through regression analysis we estimated slope parameters of significant variables and could let the persons concerned know which variables were more important or less important and suggested which variables should specially be strengthened for higher ranking of the overall evaluation and for acquisition of the excellent grade, A.

• 대한안전경영과학회지
• /
• 제26권3호
• /
• pp.101-106
• /
• 2024

The health and working conditions of employees have become increasingly important issues in modern society. In recent years, there has been a continuous rise in problems related to the deterioration of workers' alth, which seriously affects their safety and overall quality of life. Although existing research has investigated various factors affecting workers' health and working conditions, there is still a lack of studies that scientifically analyze and identify key variables from the vast number of factors. This study employs the Lasso (Least Absolute Shrinkage and Selection Operator) technique to mathematically analyze the key variables influencing workers' health status and satisfaction with their working environment. Lasso is a technique used in machine learning to identify a small number of variables that impact the dependent variable among a large set of variables, thereby reducing model complexity and improving predictive accuracy. The results of the study can be utilized in efficiently improving workers' health and working environments by focusing on a smaller set of impactful variables.

• Journal of the Korean Statistical Society
• /
• 제33권3호
• /
• pp.271-282
• /
• 2004

We discuss complete convergence of weighted sums for arrays of ø-mixing random variables. As application, we obtain the complete convergence of moving average processes for ø-mixing random variables. The result of Baum and Katz (1965) as well as the result of Li et al. (1992) on iid case are extended to ø-mixing setting.

• Journal of the Korean Statistical Society
• /
• 제34권4호
• /
• pp.263-272
• /
• 2005

Let {$X_n,\;n{\ge}1$} be a sequence of negatively associated random variables which are dominated randomly by another random variable. We discuss the limit properties of weighted sums ${\Sigma}^n_{i=1}a_{ni}X_i$ under some appropriate conditions, where {$a_{ni},\;1{\le}\;i\;{\le}\;n,\;n\;{\ge}\;1$} is an array of constants. As corollary, the results of Bai and Cheng (2000) and Sung (2001) are extended from the i.i.d. case to not necessarily identically distributed negatively associated setting. The corresponding results of Chow and Lai (1973) also are extended.

• 한국수학교육학회지시리즈B:순수및응용수학
• /
• 제31권1호
• /
• pp.57-81
• /
• 2024

We introduce a new degree reduction method for homogeneous symmetric polynomials on binary variables that generalizes the conventional degree reduction methods on monomials introduced by Freedman and Ishikawa. We also design an degree reduction algorithm for general polynomials on binary variables, simulated on the graph coloring problem for random graphs, and compared the results with the conventional methods. The simulated results show that our new method produces reduced quadratic polynomials that contains less variables than the reduced quadratic polynomials produced by the conventional methods.

• Tribology and Lubricants
• /
• 제38권1호
• /
• pp.1-7
• /
• 2022

Chemical-mechanical polishing (CMP) process is a semiconductor process that planarizes a wafer surface using mechanical friction between a polishing pad and a substrate surface during a specific chemical reaction. During the CMP process, polishing pad conditioning is applied to prevent the rapid degradation of the polishing quality caused by polishing pad glazing through repeated material removal processes. However, during the conditioning process, uneven wear on the polishing pad is inevitable because the disk on which diamond particles are electrodeposited is used. Therefore, the abrasion of the polishing pad should be considered not only for the variables during the conditioning process but also when designing the CMP conditioning system. In this study, three design variables of the conditioning system were analyzed, and the effect on the pad wear profile during conditioning was investigated. The three design variables considered in this study were the length of the conditioner arm, diameter of the conditioner disk, and distance between centers. The Taguchi method was used for the experimental design. The effect of the three design variables on pad wear and uniformity was assessed, and new variables used in conditioning system design were proposed.

• 수산해양교육연구
• /
• 제29권3호
• /
• pp.732-745
• /
• 2017

The purpose of this study was to analyze the influence of collective self-esteem on teachers' job stress. For the study, 223 teachers in elementary school, junior high school, and high school were selected and data on teachers' background variables, collective self-esteem, and job stress were collected. The relationship between variables was analysed by using multiple regression analysis. The results were as follows: The job stress of teacher was significantly different according to teacher's gender, school level, and career experience. When the influence of background variables was controlled and multiple regression analysis was performed, the group self-esteem of teachers had a great influence on mitigating the perception of job stress. Particularly, it was shown that private collective self-esteem and importance to identity are the most important factors in mitigating job stress perception among the sub-variables of collective self-esteem. These results showed that it is an important task for our society to find an effective way to elevate the teacher's collective self-esteem. Teachers' collective self-esteem is expected to play a positive role not only in mitigating job stress but also in enhancing self-fulfillment and personal self-esteem of teacher.

• 대한가정학회지
• /
• 제35권6호
• /
• pp.71-86
• /
• 1997

The purpose of this study was to investigate the effect of parent-bonding patterns & environmental variables on rural high school students' self-differentiationf. The subjects were 600 rural high school students in the 2nd grade. The rural high school students' self-differentiation, according to parent-bonding style & environmental variables were researched with DOSS(Differentiation of self Scale) PBI(Parent-bonding Instrument) & environmental Scale. The data were analyzed by frequency & Reliability, ANOVA (scheffe'test) & Regression analysis of SPSS PC+ program. The main results were as follows. First, In P-B patterns generally frequencies of mother's care were higher than father's But in overprotection frequencies of father's were higher. Degree of rural highschool students' self-differentiation was on the level of 2.84 Second, According to parent-bonding patterns, the difference of self-differentiation were as follow. In P-B patters, father and mother on each level of P<.001 brought differences in the self-differentiation of the 4 sub-divisions Third, In short the effect on self-differentiation showed that in family environmental variables, parent's scolding, gender, family atmosphere & health conditions are crucial variables in self-differentiation and in P-B patterns, father's overprotection, mother's care, father's care are all influential on self-differentiation.

• 대한가정학회지
• /
• 제48권2호
• /
• pp.95-111
• /
• 2010

We examined gender-specific behaviors in children and classified types of bullying behavior among 1,181 fifth and sixth grade elementary schools students. Differences were identified in individual variables, family environment variables, and school environment variables. Furthermore the behavioral tendencies of those variables towards bullying were also investigated. Collected data were subjected to descriptive and comparative statistical analysis using the SPSS program(Ver 15.0). Results showed that tendency towards bullying was gender specific. Bullying behavior, reinforcing behavior, assistant behavior, and onlooking behavior in boys were influenced by individual factors only. However, defending behavior in boys was influenced by individual and school factors, while victimizing behavior was influenced by individual and family factors. In girls, onlooking behavior was only influenced by individual factors, while reinforcing behavior was influenced by individual and family factors. Bullying behavior, defending behavior, assistant behavior, and victimizing behavior in girls were influenced by individual, family, and school factors.