• Title, Summary, Keyword: intervening variable

Search Result 64, Processing Time 0.038 seconds

A study on decision tree creation using intervening variable (매개 변수를 이용한 의사결정나무 생성에 관한 연구)

  • Cho, Kwang-Hyun;Park, Hee-Chang
    • Journal of the Korean Data and Information Science Society
    • /
    • v.22 no.4
    • /
    • pp.671-678
    • /
    • 2011
  • Data mining searches for interesting relationships among items in a given database. The methods of data mining are decision tree, association rules, clustering, neural network and so on. The decision tree approach is most useful in classification problems and to divide the search space into rectangular regions. Decision tree algorithms are used extensively for data mining in many domains such as retail target marketing, customer classification, etc. When create decision tree model, complicated model by standard of model creation and number of input variable is produced. Specially, there is difficulty in model creation and analysis in case of there are a lot of numbers of input variable. In this study, we study on decision tree using intervening variable. We apply to actuality data to suggest method that remove unnecessary input variable for created model and search the efficiency.

Design Optimization Using Two-Point Diagonal Quadratic Approximation (이점 대각 이차 근사화 기법을 적용한 최적설계)

  • Choe, Dong-Hun;Kim, Min-Su;Kim, Jong-Rip;Jeon, Jae-Yeong
    • Transactions of the Korean Society of Mechanical Engineers A
    • /
    • v.25 no.9
    • /
    • pp.1423-1431
    • /
    • 2001
  • Based on the exponential intervening variable, a new two-point approximation method is presented. This introduces the shifting level into each exponential intervening variable to avoid the lack of def inition of the conventional exponential intervening variables due to zero-or negative-valued design variables. Then a new quadratic approximation whose Hessian matrix has only diagonal elements of different values is proposed in terms of these intervening variables. These diagonal elements are determined in a closed form that corrects the typical error in the approximate gradient of the TANA series due to the lack of definition of exponential type intervening variables and their incomplete second-order terms. Also, a correction coefficient is multiplied to the pre-determined quadratic term to match the value of approximate function with that of the previous point. Finally, in order to show the numerical performance of the proposed method, a sequential approximate optimizer is developed and applied to solve six typical design problems. These optimization results are compared with those of TANA-3. These comparisons show that the proposed method gives more efficient and reliable results than TANA-3.

Efficient Mechanical System Optimization Using Two-Point Diagonal Quadratic Approximation in the Nonlinear Intervening Variable Space

  • Park, Dong-Hoon;Kim, Min-Soo;Kim, Jong-Rip;Jeon, Jae-Young
    • Journal of Mechanical Science and Technology
    • /
    • v.15 no.9
    • /
    • pp.1257-1267
    • /
    • 2001
  • For efficient mechanical system optimization, a new two-point approximation method is presented. Unlike the conventional two-point approximation methods such as TPEA, TANA, TANA-1, TANA-2 and TANA-3, this introduces the shifting level into each exponential intervening variable to avoid the lack of definition of the conventional exponential intervening variables due to zero-or negative-valued design variables. Then a new quadratic approximation whose Hessian matrix has only diagonal elements of different values is proposed in terms of these shifted exponential intervening variables. These diagonal elements are determined in a closed form that corrects the typical error in the approximate gradient of the TANA series due to the lack of definition of exponential type intervening variables and their incomplete second-order terms. Also, a correction coefficient is multiplied to the pre-determined quadratic term to match the value of approximate function with that of the previous point. Finally, in order to show the numerical performance of the proposed method, a sequential approximate optimizer is developed and applied to solve six typical design problems. These optimization results are compared with those of TANA-3. These comparisons show that the proposed method gives more efficient and reliable results than TANA-3.

  • PDF

Design Optimization Using Two-Point Diagonal Quadratic Approximation(TDQA) (이점 대각 이차 근사화(TDQA) 기법을 적용한 최적설계)

  • Kim, Min-Soo;Kim, Jong-Rip;Choi, Dong-Hoon
    • Proceedings of the KSME Conference
    • /
    • /
    • pp.386-391
    • /
    • 2001
  • This paper presents a new two-point approximation method based on the exponential intervening variable. To avoid the lack of definition of the conventional exponential intervening variables due to zero- or negative-valued design variables the shifting level into each exponential intervening variable is introduced. Then a new quadratic approximation, whose Hessian matrix has only diagonal elements of different values, is proposed in terms of these intervening variables. These diagonal elements are computed in a closed form, which correct the typical error in the approximate gradient of the TANA series due to the lack of definition of exponential type intervening variables and their incomplete second-order terms. Also, a correction coefficient is multiplied to the pre-determined quadratic term to match the value of approximate function with that of the original function at the previous point. Finally, the authors developed a sequential approximate optimizer, solved several typical design problems used in the literature and compared these optimization results with those of TANA-3. These comparisons show that the proposed method gives more efficient and reliable results than TANA-3.

  • PDF

A study on insignificant rules discovery in association rule mining (연관성규칙에서 의미 없는 규칙의 발견에 관한 연구)

  • Cho, Kwang-Hyun;Park, Hee-Chang
    • Journal of the Korean Data and Information Science Society
    • /
    • v.22 no.1
    • /
    • pp.81-88
    • /
    • 2011
  • Association rule mining searches for interesting relationships among items in a given database. There are three primary quality measures for association rule, support and confidence and lift. In order to improve the efficiency of existing mining algorithms, constraints were applied during the mining process to generate only those association rules that are interesting to users instead of all the association rules. When we create relation rule, we can often find a lot of rules. This can find rule that direct relativity by intervening variable does not exist. In this study we try to discovery an insignificant rule in association rules by intervening variable. Result of this study can understand relativity about rule that is created in relation rule more exactly.

The Consumer Complaining Behavior according to Consumer Problems -in case of household electric goods- (소비자문제경험에 따른 대응행동 -가전제품을 중심으로-)

  • 이승신
    • Journal of the Korean Home Economics Association
    • /
    • v.33 no.3
    • /
    • pp.179-191
    • /
    • 1995
  • The purpose of this study was to investgate consumer problems which were severe for consumers and their complaining behavior. This research had a purpose to enhance the ability of consumer protection by the well-developed consumer education. The subjects were 342 homewives in Seoul, Feb. 1995. The major results were as follows : Younger, higher-educated, small sized family and white-collor and professional occupation homewives experienced more consumer problems. Also theses homewives with higher level of consumer attitude, information search and more consumer problems had more complaining behavior. For the causal relation of consumer problems experience, several independent variables and consumer complaining behavior, the path analysis was used. Only the consumer problem experience which was a intervening variable had direct affect to the consumer complaining behavior. Education level, number of family, consumer knowledge had inderect affect through the intervening variable to the consumer complaining behavior.

  • PDF

A Study on Energy Conservation behaviors of Consumers and Related factors (소비자의 에너지 소비 절약 행동 및 관련요인에 관한 연구 -서울시의 주부를 중심으 로-)

  • 백경미;이기춘
    • Journal of Korean Home Management Association
    • /
    • v.5 no.2
    • /
    • pp.29-44
    • /
    • 1987
  • The objectives of this study were; (1)to determine if socio-demographic variables influence energy conservation attitude, energy conservation knowledge, energy conservation behavior of consumer, (2) to determine if energy conservation attitude and energy conservation knowledge influence energy conservation behavior of consumer, (3) to examine the causal relationships among variables which influence energy conservation behavior of consumer. The model for this study included following variables; (1) socio-demographic variables, which are the independent variables; (2) energy conservation behavior of consumer, which is a dependent variable; (3) energy conservation attitude of consumer, which is an intervening variable ;(4) energy conservation knowledge of consumer, which is an intervening variable. The data used in this study included 502 homemakers living in Seoul. Statistics used for data analysis were ANOVA, Pearson's Correlation and Path Analysis. The major findings were follows; (1) Energy conservation attitude and energy conservation behavior difference significantly according to the level of family income among socio-demographic variables such a age, educational level and family income. Energy conservation knowledge differenced significantly educational level, and family income. Energy conservation knowledge difference significantly educational level. (2) Energy conservation behavior differed significantly according to energy conservation attitude and energy conservation knowledge. (3) Energy conservation attitude and energy conservation knowledge had positive effects on energy conservation behavior directly and indirectly. (4) Family income had negative effects on energy conservation behavior directly and indirectly through energy conservation attitude. Education had positive effect on energy conservation behavior indirectly through energy conservation knowledge.

  • PDF

A study on association rule creation by marginally conditional variables (주변 조건부 변수에 의한 연관성 규칙 생성에 관한 연구)

  • Cho, Kwang-Hyun;Park, Hee-Chang
    • Journal of the Korean Data and Information Science Society
    • /
    • v.23 no.1
    • /
    • pp.121-129
    • /
    • 2012
  • Association rule mining searches for interesting relationships among items in a given database. Currently, study of the constraint-based association rules are underway by many researchers. When we create relation rule, we can often find a lot of rules. Of this rules, we can find rule that direct relativity by marginally conditional variables (intervening variable, external variable) does not exist. In such a case, this association rule can be considered insignificant. In this study, we want to study for association rules creation using marginally conditional variable. The result of this study can find meaningless association rules. Also, we can understand more exactly the relationships between variables.

A study on decision tree creation using marginally conditional variables (주변조건부 변수를 이용한 의사결정나무모형 생성에 관한 연구)

  • Cho, Kwang-Hyun;Park, Hee-Chang
    • Journal of the Korean Data and Information Science Society
    • /
    • v.23 no.2
    • /
    • pp.299-307
    • /
    • 2012
  • Data mining is a method of searching for an interesting relationship among items in a given database. The decision tree is a typical algorithm of data mining. The decision tree is the method that classifies or predicts a group as some subgroups. In general, when researchers create a decision tree model, the generated model can be complicated by the standard of model creation and the number of input variables. In particular, if the decision trees have a large number of input variables in a model, the generated models can be complex and difficult to analyze model. When creating the decision tree model, if there are marginally conditional variables (intervening variables, external variables) in the input variables, it is not directly relevant. In this study, we suggest the method of creating a decision tree using marginally conditional variables and apply to actual data to search for efficiency.

A study on the visual preference prediction of interiors (실내공간에서의 시각적 선호도 예측에 관한 연구)

  • 노정실;김유일
    • Journal of the Korean Institute of Landscape Architecture
    • /
    • v.26 no.2
    • /
    • pp.269-282
    • /
    • 1998
  • The visual preference of interiors focusing on lobbies was investigated as a function of six predictor variable on the base of the Informational Approach: complexity, coherence, mystery, spaciousness, brightness, plant. The Common Fcator Analysis of preference ratings yielded six common factors which helped to account for 22.3 percent of the variance in preference response to the scene. Among these factors, the factor defined as 'bright with many plants' was the most preferred and the factor defined as 'simple and closed' was the least preferred. The environmental attributes reflected in six groups of scenes were colour, resting place, window and the six predictors. In the commercial building scenes, complexity, spaciousness, coherence, brightness and mystery out of six predictors accounted for 74 percent of preference variance as the significant contributors. In the business building scenes, three predictors which are brightness, complexity, spaciousness accounted for 84 percent of preference variance. 'The amount of plant' not only influenced the preference indirectly through the intervening variable, complexity, but also was moderately correlated with brightness. The overall pattern of the resulted confirmed the usefulness of the Informational Approach to predict the preference in interiors focusing on lobbies.

  • PDF