탐색적 요인분석: 어떻게 달라지나?

이순묵,윤창영,이민형,정선호 한국심리학회 2016 한국심리학회지 일반 Vol.35 No.1

In the present study new developments in EFA(Exploratory Factor Analysis) that have occurred at the turn of the 21th century are discussed. New guidelines and an analysis of real data following the guidelines are given with practical comments. First, in a process of determining the number of factors, MRFA (minimum rank factor analysis) is recommended as the best method of Parallel Analysis (PA) instead of Horn's method (1965) and PA-PAFA (parallel analysis in principal axis factor analysis). Various fit indices such as CFI, RMSEA, and etc. allow us to consider “various psychometric criteria” before determining the number of factors as the indices are less sensitive to sample sizes than the conventional statistic. In addition estimation methods that are applicable to categorical data (dichotomous or polytomous) have been developed so that item factor analysis can be readily performed for categorical data. Second, in a process of rotating factor structures, “simple” oblique structure can be easily computed just by minimzing the value of complexity function, and a “partially specified target” rotaion is also available adopting a target matrix whose elements are partially hypothesized by a researcher. Finally, in a process of interpreting factor structures, ESEM (Exploratory Structural Equation Modeling) will get prevalence in the near future as it allows us to free correlations between unique factors (measurement errors) and can produce more practical and interpretable factor stucutures. New guidelines on EFA are described in detail in the later part of this paper. 본 연구에서는 21세기 전후하여 새롭게 제안된 탐색적 요인분석(EFA)에 대한 지침들을 정리하고 실제 자료의 분석 예를 제시하였다. 대략적인 요인수효를 결정하기 위한 발견법(heuristics)의 내용가운데 평행성 분석에 대한 평가가 정리되었고 랜덤자료에서의 고유치로서 기존의 Horn(1965) 방식이나 주축요인방식이 아닌, 최소계수요인 방식(MRFA: minimum rank factor analysis)에서의 고유치가 권고된다. 요인수효 결정을 위한 추론적 접근에서 합치도의 참조는 카이제곱 검증뿐만 아니라 표본 크기에 영향을 덜 받는 다양한 판단적 합치도(예: CFI, RMSEA 등)를 함께 참조할 수 있고 이로 인해 요인수효 결정에서 “다양한 정보의 종합적 사용”이 가능해졌다. 여기에 서열자료 분석에 사용될 수 있는 추정법이 개발되면서, 문항점수들을 연속변수에 준하는 것으로 보고 피어슨상관을 구하여 고전적 요인분석을 하는 관행을 벗어나 문항의 범주별 반응형태를 반영하는 문항요인분석이 현실화되었다. 요인구조의 회전에 있어서는 사각구조의 추정이 용이해졌고, 임의적인 파라메터의 설정 없이 복잡도 함수만을 최소화함으로써 단순구조를 추구할 수 있게 되었다. 또한 탐색과정에서 연구자의 내용적 판단을 반영하는 목표행렬을 주고 그 방향을 따르도록 회전하는 부분제약 목표회전의 사용이 가능해져 이전의 기계적인 회전을 벗어나게 되었다. 요인구조의 해석 가능성에서 가장 큰 변화로 볼 수 있는 것은 측정오차 간 상관을 허용하는 탐색적 구조방정식 모형(ESEM: Exploratory Structural Equation Modeling)이 개발되어, EFA를 할 때 측정오차 간 상관이 없다는 종래의 강한 가정을 완화시키면서 현실적이고 해석 가능한 구조를 산출하게 되었다. 실제 자료의 분석 예시에서는 탐색적 요인분석에서 새로운 지침들이 어떻게 활용되고 있는지를 상세히 설명하고 있다.

유통과학분야에서 탐색적 연구를 위한 요인분석

임명성 한국유통과학회 2015 유통과학연구 Vol.13 No.9

Purpose – This paper aims to provide a step-by-step approach to factor analytic procedures, such as principal component analysis (PCA) and exploratory factor analysis (EFA), and to offer a guideline for factor analysis. Authors have argued that the results of PCA and EFA are substantially similar. Additionally, they assert that PCA is a more appropriate technique for factor analysis because PCA produces easily interpreted results that are likely to be the basis of better decisions. For these reasons, many researchers have used PCA as a technique instead of EFA. However, these techniques are clearly different. PCA should be used for data reduction. On the other hand, EFA has been tailored to identify any underlying factor structure, a set of measured variables that cause the manifest variables to covary. Thus, it is needed for a guideline and for procedures to use in factor analysis. To date, however, these two techniques have been indiscriminately misused. Research design, data, and methodology – This research conducted a literature review. For this, we summarized the meaningful and consistent arguments and drew up guidelines and suggested procedures for rigorous EFA. Results – PCA can be used instead of common factor analysis when all measured variables have high communality. However, common factor analysis is recommended for EFA. First, researchers should evaluate the sample size and check for sampling adequacy before conducting factor analysis. If these conditions are not satisfied, then the next steps cannot be followed. Sample size must be at least 100 with communality above 0.5 and a minimum subject to item ratio of at least 5:1, with a minimum of five items in EFA. Next, Bartlett's sphericity test and the Kaiser-Mayer-Olkin (KMO) measure should be assessed for sampling adequacy. The chi-square value for Bartlett's test should be significant. In addition, a KMO of more than 0.8 is recommended. The next step is to conduct a factor analysis. The analysis is composed of three stages. The first stage determines a rotation technique. Generally, ML or PAFwill suggest to researchers the best results. Selection of one of the two techniques heavily hinges on data normality. ML requires normally distributed data; on the other hand, PAF does not. The second step is associated with determining the number of factors to retain in the EFA. The best way to determine the number of factors to retain is to apply three methods including eigenvalues greater than 1.0, the scree plot test, and the variance extracted. The last step is to select one of two rotation methods: orthogonal or oblique. If the research suggests some variables that are correlated to each other, then the oblique method should be selected for factor rotation because the method assumes all factors are correlated in the research. If not, the orthogonal method is possible for factor rotation. Conclusions – Recommendations are offered for the best factor analytic practice for empirical research.

오픈소스 GIS를 활용한 공간분석 프레임워크 및 실증적 연구 -부동산 관련 데이터를 사례로-

김정희 ( Kim Jung Hee ) 한국사진지리학회 2020 한국사진지리학회지 Vol.30 No.4

This study is designed to provide the spatial analysis framework by using the open source GIS as the decision-making supporting means for the real estate issues and make the empirical analysis which can support it. For this purpose, the real estate issues which would be mainly handled in the spatial analysis applied with GIS would be derived and a series of analysis process framework would be suggested from the construction of data to the spatial analysis techniques such as exploratory spatial data analysis and model-driven spatial data analysis. Based on this, the characteristics and functions of the open source GIS which is proper for the analysis would be found and the analysis method would be searched before the empirical analysis would be performed by using the data of the rent trend for the commercial real estate located in Seoul. This study would focus on the suggestion of guideline for a kind of spatial analysis and how to use it.

체육학에서 요인분석의 통계적 타당화: 탐색적, 확인적 요인분석을 중심으로

정연택,최연재 한국코칭능력개발원 2023 코칭능력개발지 Vol.25 No.5

요인분석은 체육학에서 널리 이용되는 통계기법이며 사용 빈도가 높다. 요인분석은 변수를 축소하거나 제거하여 측정하는 항목의 타당성을 검증한다. 특히, 요인분석은 측정된 자료를 이용하여 변수 생성 등의 목적을 가진다. 따라서 이 연구는 체육학에서 널리 사용하고 있는 분석의 개념과 탐색적, 확인적 요인분석에 대하여 이해하고자 하였다. 1장에서는 요인분석의 특성과 중요성에 관하여 서술하였으며, 2장은 요인분석의 개념과 기본과정에 관한 서술이 이루어졌다. 3장에서는 분석 절차 중 탐색적, 확인적 요인분석의 중요도를 설명하였다. 4장에서는 앞서 설명된 요인분석을 바탕으로 논의를 하였으며, 5장은 통계분석의 합리적 방안을 제시하였다. Factor analysis is a statistical technique widely used in physical education and is frequently used. Factor analysis verifies the validity of items to be measured by reducing or removing variables. In particular, factor analysis has the purpose of creating variables using measured data. Therefore, this study tried to understand the concept of factor analysis widely used in physical education and exploratory and confirmatory factor analysis. Chapter 1. describes the characteristics and importance of factor analysis, and Chapter 2. describes the concept and basic process of factor analysis. Chapter 3. explained the importance of exploratory factor analysis and confirmatory factor analysis among the procedures of factor analysis. In Chapter 4. discussions were made based on the factor analysis described above, and in Chapter 5. rational methods for statistical analysis were presented.

체육수업 재미거리 질문지(PCESQM) 타당성 및 신뢰성 검증

박준성(Jun Sung Park) 한국사회체육학회 2009 한국사회체육학회지 Vol.0 No.37

This paper verified the validity of the structure and reliability whether the questionnaire on the source of enjoyment in physical education classes for middle school students developed by Chang-sub Lee and Sang-woo Nam(2003) can effectively measure the fun factors in the physical education classes for the male and female students(592) of middle schools in S city in 2008. As a result of the item analysis and the reliability analysis, since questions with a mean over 4.5 and questions with the skewness and kurtosis values over ±2.0 did not appear, the results of this paper showed to have an appropriate level of the mean, standard deviation, skewness, kurtosis, and reliability. Furthermore, as the results of the correlation matrix between questions, show that since more than 50% of the questions have a correlation over ±3.0 between each question, the question of this paper are judged to be adequate questions for the factor analysis. Moreover, in the exploratory factor analysis, the maintenance of physical and mental health factor was developed into questions number 2, 3, 5, 6, 7, sense of accomplishment factor to questions number 1, 4, 12, the easiness of the class factor to questions number 9, 10, 15, the freedom of the class factor to questions number 13, 14, 16, and the sociability cultivation factor to questions number 8, 11, 17 that are the sub-factor of the questionnaire on the source of fun in physical education classes for middle school students developed by Chang-sub Lee and Sang-woo Nam(2003). Also in the exploratory factor analysis of this paper, as the questions related to factor 1 showed to be questions number 6, 7, 5, 2, 3, factor 2(1, 4, 12), factor 3(10, 9, 15), factor 4(14, 13, 16), and factor 5(17, 8, 11), it indicates that is composed of the five sub-factors developed by Chang-sub Lee and Sang-woo Nam(2003). For the confirmatory factor analysis that assesses the fitness of the model, as the Q value was 2.004, GFI .910, AGFI .897, CFI .884, RMSEA .71, and RMR .51, it is judged that the model is appropriate. As a result of evaluating the fitness of the model through the item analysis, correlation analysis, exploratory factor analysis, and the confirmatory factor analysis as mentioned above, the questionnaire on the source of enjoyment in physical education classes for the middle school students developed by Chang-sub Lee and Sang-woo Nam(2003) showed to be an appropriate questionnaire in measuring the enjoyment factors for the male and female students of middle schools in S city.

데이터 탐색을 활용한 딥러닝 기반 제천 지역 산사태 취약성 분석

안상아,이정현,박혁진 대한지질공학회 2023 지질공학 Vol.33 No.4

데이터 탐색은 수집한 데이터를 다양한 각도에서 관찰 및 이해하는 과정으로 데이터 구조 및 특성 분석을 통해 데이터의 분포와 상관관계를 파악하는 과정이다. 일반적으로 산사태는 다양한 인자들에 의해 유발되고 발생 지역에 따라 유발 인자들이 미치는 영향이 상이하기 때문에 산사태 취약성 분석 이전에 데이터 탐색을 통해 유발 인자 사이의상관관계를 파악하고 특징적인 유발 인자를 선별한다면 효과적인 분석을 수행할 수 있다. 따라서 본 연구는 데이터 탐색이 예측 모델의 성능에 미치는 결과를 확인하기 위해 두 단계에 걸친 데이터 탐색을 수행하여 인자를 선별하고, 선별된 유발 인자들 사이의 조합과 23개의 전체 유발 인자 조합을 활용하여 딥러닝 기반의 산사태 취약성 분석을 진행하였다. 데이터 탐색 과정에서는 Pearson 상관계수 heat map과 random forest의 인자 중요도 histogram을 활용하였으며, 딥러닝 기반 산사태 취약성 분석 결과의 정확도는 분석을 통해 획득한 산사태 취약 지수 값을 이용해 제작한 산사태 취약성 지도를 confusion matrix 기반의 정확도 검증 방법을 통해 분석하였다. 분석 결과, 전체 23개의 인자를 사용한 산사태 취약성 해석 결과는 55.90%의 낮은 정확도를 보였지만 한 단계의 탐색을 거쳐 선별한 13개 인자를 활용한 취약성 해석 결과는81.25%의 분석 정확도를 보였고, 두 단계 데이터 탐색을 모두 수행하여 선별된 9개의 유발 인자를 활용한 산사태 취약성 분석 결과는 92.80% 로 가장 높은 정확도를 보였다. 따라서 데이터 탐색을 통해 특징적인 유발 인자를 선별하고 분석에 활용하는 것이 산사태 취약성 분석에서 더좋은 분석 성능을 기대할 수 있음을 확인하였다. Exploratory data analysis is the process of observing and understanding data collected from various sources to identify their distributions and correlations through their structures and characterization. This process can be used to identify correlations among conditioning factors and select the most effective factors for analysis. This can help the assessment of landslide susceptibility, because landslides are usually triggered by multiple factors, and the impacts of these factors vary by region. This study compared two stages of exploratory data analysis to examine the impact of the data exploration procedure on the landslide prediction model’s performance with respect to factor selection. Deeplearning-based landslide susceptibility analysis used either a combinations of selected factors or all 23 factors. During the data exploration phase, we used a Pearson correlation coefficient heat map and a histogram of random forest feature importance. We then assessed the accuracy of our deep-learningbased analysis of landslide susceptibility using a confusion matrix. Finally, a landslide susceptibility map was generated using the landslide susceptibility index derived from the proposed analysis. The analysis revealed that using all 23 factors resulted in low accuracy (55.90%), but using the 13 factors selected in one step of exploration improved the accuracy to 81.25%. This was further improved to 92.80% using only the nine conditioning factors selected during both steps of the data exploration. Therefore, exploratory data analysis selected the conditioning factors most suitable for landslide susceptibility analysis and thereby improving the performance of the analysis.

K-SIB-R을 통한 정신지체 아동 적응행동의 이론적 구조 탐색

김은영,유구종 한국특수교육학회 2008 특수교육학연구 Vol.43 No.1

The goal of the present study was to further elucidate the theoretical factor structure of adaptive behavior indexed by K-SIB-R battery in mentally retarded children and adolescent. K-SIB-R battery is korean version of SIB-R. Exploratory factor analysis based on the eigen value criteria revealed that K-SIB-R has one factor solution, which is consistent with the work of Paik et al. (2005)’confirmatory factory analysis. However, factor analysis based on the scree plot criteria suggested that three factor model should be considered as an alternative. The needs of subsequent exploratory factor analysis of the data which has been already well analyzed using Confirmatory factor analysis is the pitfall of 'confirmation bias' underlying structure equation modeling (SEM), which is a tendency in SEM to interpret the data in a way that confirms in the working hypothesis of the study. Therefore, caution should always be taken when making claims of causality: good fit by a model consistent with one causal hypothesis does not rule out equally good fit by another model consistent with a different causal hypothesis. it is necessary to test rival hypotheses produced by exploratory factor analysis in relevant interpreting existing data. The results of the present study is in line with the previous findings that suggest the correlation of the cognitive ability with the adaptive behavior skill. 본 연구에서는 정신지체 아동을 위한 적응행동 평가 도구(Scales of Independent Behavior-Revised: SIB-R)의 행동 척도를 국내에서 표준화한 한국판 적응행동 평가 도구(Korean SIB-R: K-SIB-R)를 사용하여 적응행동의 이론적 구조에 대한 탐색적 요인 분석을 시행하고 다양한 모형을 탐색하였다. 서울, 경기, 충남, 전남의 6개 학교 61명의 정신지체 아동들이 연구에 참여하였다. 결과, K-SIB-R은 단일요인모델 뿐만 아니라 3요인 모델로도 설명할 수 있음을 나타내었다. 단일요인모델은 선행연구(백은희, 이병인, 조수제, 2005)에서 행해진 확인적 요인분석의 결과와 일치하였다. 확인적 요인분석으로 이론적 요인규명이 끝난 것으로 보이는 연구에 다시 탐색적 요인 분석 필요성이 제기되는 것은 구조방정식의 ‘확증편향성(confirmation bias)’ 때문이다. 따라서, 제안된 모델을 검정하는 과정에서 상이한 구조관계를 가정하고 있는 경쟁모델들을 발견하여 검정하는 과정은 필수적이다. 한편 본 연구에서는 탐색된 요인으로 분산분석을 실시하여 인지기능의 차이가 적응행동의 발달 과정에 일관된 형식(pattern)으로 작용하고 있음을 발견하였다.

퇴직준비 예측모형에 있어 척도의 타당화 연구 : 탐색적 및 확인적 요인분석의 통합 적용

박철민,권재영 한국산업경제학회 2019 산업경제연구 Vol.32 No.6

This study aims at suggesting stable scales pertaining to the retirement preparation model through the research of previous studies. In doing so, this study applied a questionnaire methods in research of public employees in Busan areas. Collecting data through 389 employees in public sector working in the Busan region for empirical research are implemented. Used techniques in analyzing data were exploratory & confirmatory factor analysis mainly. SPSS statistical program and Amos programs were used in this analysis. In attaining research goals, frist, the study applied refining works of a series of scales through the exploratory factor analysis. Second, confirmatory factor analysis was applied at the refined scales. In this stage, lower level of scales in their explaining power were removed. Totally through 7 rounds of confirmatory factor analysis, stable scales of retirement preparation model were suggested. Finally, research model of this study was tested through these stable scales of retirement preparation. 이 연구는 조직 구성원들의 퇴직 준비 예측모형에서 이를 구성하는 척도의 타당도를 검정하여 향후 관련 연구에 있어 이용 가능한 안정적인 측정도구를 제시하는 데 목적을 둔다. 이를 위해 부산지역 공무원을 대상으로 자료를 수집․분석하였고, 분석방법은 탐색적 및 확인적 요인분석을 주로 이용하였다. 분석 프로그램은 SPSS와 구조방정식모델(SEM with AMOS)을 사용하였다. 연구 목적을 달성하기 위해 첫째, 탐색적인 요인분석법(EFA)을 통해서 측정변수에 대한 정제 작업을 실시하였다. 둘째, 이들 정제된 변수를 대상으로 확인적인 요인분석법(CFA)을 실시하였다. 이 과정에서 퇴직준비 관계 예측을 위한 척도 모델의 적합도를 높이기 위해 설명력이 낮은 측정변인을 수차례의 단계를 거쳐 소거하면서 확인적인 요인분석법을 모두 7번 실시하여, 최종적인 척도를 제시하고 연구모형을 검증하였다.

The Exploratory Analysis for Spam Mail Data Using Correspondence Analysis

신양규 한국데이터정보과학회 2005 한국데이터정보과학회지 Vol.16 No.4

The number of electronic mail(E-mail) has been increased dramatically as a result of expanding internet and information technology. Although there are many conveniences of E-mail in the bright side, some serious problems occur because of E-mail in its dark side. One of the problems is spam-mail which is unsolicited mail and also called bulk mail. This paper presents a set of patterns of spam-mail occurrences within a week using the correspondence analysis. The correspondence analysis is an exploratory multivariate technique that converts data into a particular type of graphical display in which the rows and columns are depicted as points. One of the meaningful patterns is a great increment of adult and phishing related spam-mails at weekends so any spam-mail filters should be designed to cope with this pattern.

다차원 정책분석모형 활용에 관한 탐색적 연구

김제현(Je Hyun Kim) 한국지방교육경영학회 2020 지방교육경영 Vol.23 No.3

본 연구는 Cooper 외(2004) 다차원 정책분석모형을 활용한 다수의 선행연구를 대상으로, 연구자별로 정책분석모형을 활용함에 있어 다양하고도 상이한 시각 및 접근을 보이고 있는 점에 주목하여, 다차원 정책분석모형 활용에 대한 시사점을 제시해 보고자 하였다. 우선적으로 Cooper 외(2004) 다차원 분석모형을 개관하고, 해당 분석모형 활용과 관련하여 선행연구들을 탐색하였다. 다음으로, 선행연구들을 탐색한 결과를 바탕으로, Cooper 외(2004)가 제시한 ‘4가지 윤리적 장벽’ 문제(정책분석의 기술 관료적 특성; 널리 통용되는 이론적 모델의 부재; 정책 연구의 고객 중심적 성격; 윤리 분석에 따르는 심리적 위험성)와 결부시켜 정책철학적 접근의 부재, 정책패러다임 접근의 필요성, 다차원 정책분석 틀의 수립 방향을 논의하였다. 논의 결과, 첫째, 정책철학적 접근이 필요하며, ‘윤리와 사회정의’ 핵심가치 차원을 강조해야 한다. 둘째, 정책패러다임 접근이 필요하다. 셋째, 연구자 편향의 정책분석을 지양하고, 이론적 분석 틀 수립의 가이드라인이 필요하다. 연구결과를 토대로 한 제언은 다음과 같다. 첫째, 후속연구로서, Cooper 외(2004)의 다차원 정책분석모형에 관한 이론적 검증 연구가 필요하다. 둘째, Cooper 외(2004)의 다차원 정책분석에 있어서 정책철학적 접근이 필요하기에, 정책학적으로 정책철학적 접근에 대한 명확한 규명이 필요하다. This study targets a number of previous studies using the multidimensional policy analysis model of Cooper et al. (2004), and focuses on the fact that each researcher shows various and different perspectives and approaches in using the policy analysis model. I tried to present implications for the use of the model. First of all, Cooper et al. (2004) reviewed the multidimensional analysis model and searched for previous studies in relation to the use of the analysis model. Next, based on the search results of previous studies, the problem of'four ethical barriers' proposed by Cooper et al. ; Psychological risk following ethical analysis), the absence of a policy philosophy approach, the necessity of a policy paradigm approach, and the direction of establishment of a multidimensional policy analysis framework were discussed. As a result of the discussion, first, a policy philosophy approach is needed, and the core value dimension of'ethics and social justice' should be emphasized. Second, a policy paradigm approach is needed. Third, it is necessary to avoid policy analysis of researcher bias and to establish a theoretical framework for analysis. Suggestions based on the research results are as follows. First, as a follow-up study, a theoretical verification study on the multidimensional policy analysis model of Cooper et al. (2004) is needed. Second, in the multidimensional policy analysis of Cooper et al.