데이터마이닝을 활용한 인구집단 구성과 경제적 불평등 평가

민대기 ( Dai Ki Min ) 아시아.유럽미래학회 2016 유라시아연구 Vol.13 No.1

최근 인구의 고령화 현상에 의하여 연금제도의 중요성이 증가하고 있다. 본 연구에서는 경제적 불평등 완화를 목적으로 하는 연금제도의 특성을 고려할 때 불평등 수준 평가의 기초가 되는 우리나라 인구집단 구성의 적정성을 평가하고자 한다. 이를 위하여 한국노동패널조사 자료 중 60대와 70대 고령 은퇴가구를 대상으로 데이터 마이닝 기법 중 군집분석 (clustering analysis)을 수행하여 다양한 요인들 사이의 암묵적 관계에 따라 유사한 특성을 가질 것으로 예상되는 세부 인구집단을 구성한다. 군집분석은 변수의 선정, 계층적 군집분석을 이용한 군집의 수 결정, 비계층적 군집분석을 이용한 집단 구성, 인구집단별 특성 분석 등의 네 단계로 수행한다. 군집분석 과정에서 근로기간과 풀타임 근로기간 등의 직업이력이 은퇴가구의 경제 불평등 수준에 미치는 영향을 분석하기 위하여 근로이력을 군집분석을 주요 요인으로 포함하였다. 가구특성, 소득, 소비, 자산의 관점에서 군집분석 결과를 살펴보면 기존 연구에서 고려한 우리나라 인구집단의 일부 특성이 적절하지 않음을 확인하였다. 특히 고용기간과 총 고용기간에서 파트타임 고용기간이 차지하는 비중은 은퇴가구의 소득과 자산에 영향이 큼에도 불구하고 현실과 기존연구의 가정에서 다소 차이가 있음을 확인하였다. 이와 같은 연구를 통하여 우리나라 연금제도의 효과적 설계에 필요한 현실적인 인구집단의 구성과 분석을 위한 가정 설정의 기반을 제공하였다는데 의의가 있다. 인구집단 사이의 경제적 불평등에 영향을 미치는 요인을 적절하게 고려하는 것은 매우 중요하다. 경제적 불평등에 다수의 요인이 복합적으로 작용함에도 불구하고 기존 연구는 단일 요인 별로 표준 인구집단을 구성하는 문제가 있다. 본 연구에서는 엔트로피 지수 (Generalized Entropy Index; GE Index) 분해 기법을 이용하여 인구집단 사이의 불평등 수준을 평가하고 이에 영향을 주는 주요 요인을 확인하였다. 표준 인구집단 구성 및 경제적 불평등 기여 요인으로는 성별, 교육수준, 가구 구성원 수와 같은 가구특성과 함께 근로기간 및 풀타임 근로기간과 같은 직업이력을 고려하였다. 분석 결과 다수의 요인을 복합적으로 활용하는 군집분석을 이용하여 구성한 인구집단은 표준 인구집단과 비교하여 불평등 수준을 효과적으로 설명하고 있음을 확인하였다. 이는 불평등 수준을 완화하기 위한 효과적인 연금제도를 설계함에 있어 인구집단 구성을 위한 군집분성 방안의 타당성을 제시한다. The importance of public pension system recently has been growing because of an ageing population in Korea. One of the major purposes of public pension system is to reduce economic inequalities across population groups that have been mainly defined by a single variable such as age, income, sex, etc in the literature. This paper aims to evaluate how well the population groups are practically and clearly designed in the literature so as to specify economic inequalities across different groups. For this purpose, we investigate the retired household sample obtained from KLIPS (Korean Labor and Income Panel Study) and conduct the clustering analysis to organize population groups. The clustering analysis consists of four steps: variable selection, hierarchical clustering analysis to determine the number of clusters and centroid, non-hierarchical clustering analysis (we use K-means clustering analysis) and define population groups, and describing the characteristics of population groups. The clustering analysis divides the population into six subgroups that have similar characteristics. We compare the characteristics of the population groups obtained from the clustering analysis with those designed in the literature with respect to household, income, consumption and asset. From the comparison, we observe that the assumptions on population groups in the literature are somewhat different from the findings from the clustering analysis. In particular, the period of employment and the full-time period are significantly different from the literature. The clustering analysis reveals that the period of employment is approximately 31-45 years, which is shorter than the period of employment assumed in the literature. Furthermore, few research has clearly considered the full-time period, which accounts for 14%-45% of the total period of employment. We believe that the findings in this study provides meaningful information on how to practically organize population groups as a part of designing public pension system. For example, researchers are encouraged to consider the employment periods and full-time periods obtained from the data analysis when considering the significant impact of working periods on pension payment and benefits. Population should be well grouped so as to specify the difference in economic inequalities across groups. In general, various variables are jointly involved in determining the economic inequalities, but populations have been simply grouped by a single variable such as age, gender, education level, household size, etc. in the literature. This paper employs the Generalized Entropy (GE) measure of economic inequality and its decomposition to evaluate the contributions of several variables on the overall economic inequality. For this analysis, populations are first grouped by several variables such as clusters obtained from the analysis, sex, education level, period of employment and household size. Then, we measure the inequaltiy index and its decomposition for monthly income, monthly consumption and asset. Grouping populations using the clustering analysis specifies the economic inequality more precisely provides better measure of similarity within a group and difference between groups. This result supports the effectiveness of the clustering analysis in grouping populations.

Entropy Analysis in Foreign Exchange Markets and Economic Crisis

Jin-Gi Ha,임규빈,김승환,정우성 한국물리학회 2012 THE JOURNAL OF THE KOREAN PHYSICAL SOCIETY Vol.61 No.4

We investigate the relative market efficiency in 11 foreign exchange markets by using the Lempel- Ziv (LZ) complexity algorithm and several entropy values such as the Shannon entropy, the approximate entropy, and the sample entropy. With daily data in 11 foreign exchange markets from Jan. 2000 to Sep. 2011, we observe that mature markets have higher LZ complexities and entropy values than emerging markets. Furthermore, with sliding time windows, we also investigate the temporal evolutions of those entropies from Jan. 1994 to Sep. 2011, and we find that, after an economic crisis, the approximate entropy and the sample entropy of mature markets such as Japan, Europe and the United Kingdom suddenly become lower.

한국 정부조직의 복잡성과 무질서함의 제도적 기원:‘청’(廳) 조직에 대한 ‘제도적 엔트로피’ 관점의 제도분석

권향원 서울행정학회 2023 한국사회와 행정연구 Vol.34 No.3

본 연구는 “한국 청(廳) 조직의 구조적 복잡성과 제도적 무질서함의 맥락적 기원을 경험적으로 탐구”하였다. 이를 위해 신제도주의의 ‘제도적 엔트로피’(institutional entropy)를 중심개념으로 삼아, 정부조직 제도가 초기에는 질서정연하게 자리잡혀 있었으나, 어떻게 시간을 거듭하며 무질서하고 복잡하게 변화하게 되었는지를 역사적 맥락에서 분석하였다. 분석결과 청 조직은 건국초기에는 ‘집행・관리 기구’라는 뚜렷한 권한과 기능으로 출발하였으나, 경제성장에 따라 산업부문이 성장함에 따라 이를 지원하기 위한 ‘자율・사업 기구’의 속성이 추가로 탑재되었음이 확인되었다. 이후 행정국가화에 따라 정부부문이 확대되면서 부처 간의 칸막이와 관할권을 아우르는 ‘기획・조정 기구’를 지닌 기구에 대한 필요성이 대두되었고, 이 과정에서 청 조직에는 애초의 ‘집행・관리 기구’와는 상반된 기획과 조정의 속성까지 추가로 부여되었다. 이러한 과정에서 청 조직은 초기의 본래적 정의와는 결이 다른 예외들이 개념적으로 추가되었으며, 이러한 메커니즘과 동인으로 인하여 ‘제도적 엔트로피’가 지속적으로 상승하였던 것으로 해석된다. 결론적으로 연구자는 이러한 “지속적인 예외의 추가”가 한국 정부조직의 무질서와 비효율성의 제도적 기원에 대한 하나의 해석을 내어주는 것으로 논하였다. 이를 바탕으로 아직은 다소 생소한 ‘제도적 엔트로피’ 개념의 이론적 의의 및 분석결과의 실무적 시사점을 다루었다. This study empirically analyzed “through what mechanisms and drivers the structural complexity and institutional disorder of Korean agencies organizations were institutionalized.” In the process, through institutional analysis of historical records based on the central concept of ‘institutional entropy’ of new institutionalism, the government organization system was initially structured in an orderly manner, but disorder and complexity increased over time. and traced the driver. As a result of the analysis, in the early days of the founding of the country, the agency organization started with an ‘institutional logic’ as an ‘executive and management organization’, but as industry and economic scale grew, the attributes of an ‘autonomous and business organization’ were added, and later, the jurisdiction of individual ministries and Due to the increase in challenges across partitions, the expansion of authority and functions to a ‘planning and coordination body’ targeting all ministries was explored. In this process, institutional exceptions to the original institutional concept of the agency organization were continuously expanded and added, and it was confirmed that the ‘institutional entropy’ of the agency organizational system also increased due to these mechanisms and drivers. Based on this, the theoretical significance of ‘institutional entropy’ as a theoretical framework for diachronic institutional analysis and the practical implications of the analysis results were discussed.

기후요소의 엔트로피와 그 순변화형에 의한 기후지역 구분

박현욱 국토지리학회 2007 국토지리학회지 Vol.41 No.1

The seasonal variation and frequency of precipitation phenomenon of the Korean Peninsula over a period of 10 days during summer show strong local weather phenomena because of its mathematical location and geographic factors. The data used for this study are daily precipitation over a period of 10 days of summer during the recent fifteen years (1991-2005) at 68 stations in Korea. It is divided into four class of no rain,1-10.0mm/day, 10.1-30.0mm/day, 30.1mm over/day. The purpose of this paper is to induce the entropy of climatic elements over a period 10 days during summer, clarify the variation patterns of their space scale, and study the classification of climatic region in Korea, according to the combinations of the entropy and variation pattern, using the entropy and amplitude coefficient (Rs) calculated on basis of the information theory and the principal component analysis. The characteristics of variation pattern over a period of 10 days during summer can be chiefly divided into four categories (Rs1, Rs2, Rs3, Rs4) and the accumulated contributory rate of this is 84.1%. The variation pattern of entropy during a period of 10 days in Korea classified into 9 types from A to j. And Climatic region established by the entropy of climatic elements during summer in Korea can be classified into 14 types. 본 연구는 한반도의 수리적 위치와 지리적 요인의 영향을 명확히 반영하여 우리나라의 날씨 및 기후특성을 잘 나타내는, 하계의 탁월일기 및 강수현상의 순별 다소 {강수 없음, 소우(0.1-10.0mm/일), 중우(10.1-30.0mm/일), 대우(30.1mm 이상/일)}와 그 변화에 대해, 정보이론과 주성분 분석법을 응용하여 기후요소의 엔트로피와 주성분 진폭계수(Rs)를 추출하고 적용하여 그 공간스케일의 시간적 변동을 살핀다. 이를 토대로 한국 내 68개 관측지점 별 하계 순(1991-2005)의 기후요소의 엔트로피의 기후대표성을 구명하고 기후지역 구분을 시도한 것이다. 그 결과, 기후요소의 엔트로피의 순변화의 전형적 특징은 상위 4개의 순변화형(Rs1, Rs2, Rs3, Rs4)으로 표현되며 그 누적 기여율은 84.1%이다. 또한 한국의 기후요소의 엔트로피의 순변화형은 A-j형 까지 9개가 추출되었고 기후지역은 14개 형으로 분류되었다.

Selection of Optimal Decomposition Layer for Thresholding Denoising Using Singular Spectrum Analysis and Wavelet Entropy

Zhi Cui,Xian-pu Cui 보안공학연구지원센터 2016 International Journal of Multimedia and Ubiquitous Vol.11 No.4

To optimize the number of decomposition layers in wavelet threshold denoising for ultrasonic signals, we propose a self-adaptive algorithm of determining the number of decomposition layers based on singular spectrum analysis and wavelet entropy. First the noise-containing signals are decomposed by discrete wavelet transform. The slope of the singular value spectrum for each layer is estimated. Then the wavelet entropy over the signal subinterval is calculated for each layer. Finally the optimal number of decomposition layer is determined by combining the entropy ratio of detail coefficients to original signal and the slope of the singular value spectrum. The performance of the algorithm is evaluated using signal-to-noise ratio (SNR) and the relative error of the peak value (REPV). Experiment shows that the algorithm can self-adaptively determine the optimal number of decomposition layers and filter out the noise contained in the ultrasonic signals. It not only increases the SNR, but also preserves valuable components of the original signal.

DEA Window 분석에서 최적의 윈도우 길이 결정방법

박만희(Park Man Hee) 글로벌경영학회 2019 글로벌경영학회지 Vol.16 No.3

본 연구에서는 시계열데이터에 대한 효율성 분석방법으로 이용되는 DEA(Data Envelopment Analysis) Window 분석에서 최적 윈도우 길이를 결정할 수 있는 방법을 제시하였다. 최적 윈도우의 길이를 결정하기 위해서 효율성 값들에 대한 변별력 평가기준으로 Shannon의 엔트로피 기반중요도 지수를 이용하였으며, 실증분석을 통해 제안된 방법의 유용성을 입증하였다. 본 연구에서 제안된 최적 윈도우 길이 결정방법의 실증분석을 위하여 3개 예제가 적용되었다. 적용결과에 따르면 제안된 방법이 기존 방법보다 적용 가능성이 높은 것으로 분석되었다. Kruskal-Wallis 순위합 검정에 따르면 윈도우 길이에 따른 DMU들 간의 효율성 순위에는 차이가 없는 것으로 분석되었다. 따라서 최적 윈도우 길이 결정 알고리즘을 DEA Window분석에 이용할수 있다. 기존 연구에서는 명확한 변별력 평가기준 없이 비교대상 DMU의 수가 가장 많은 경우가 효율성값들에 대한 변별력이 높을 것이라는 가정을 하고 있다. 본 연구에서는 이런 문제를 해결하기 위해 엔트로피 기반 변별력 평가기준을 적용한 최적 윈도우의 길이를 결정할 수 있는 방법을 제안하였다. 본 연구에서 제안한 최적 윈도우 길이 결정방법을 적용한 결과에 따르면 전체 비교횟수가 가장 많은 경우보다 1회 비교횟수가 가장 많은 경우의 윈도우 길이가 최적 윈도우 길이라는 것이 입증되었다. This paper presents a method for determining the optimal window length in DEA window analysis which is used as an efficiency analysis method of time series data. To determine the optimal window length, an algorithm was presented to determine the optimal window length using Shannon’s Entropy-based importance weights as a criterion for discriminating against efficiency values, and empirical analysis proved the usefulness of the proposed method. In this paper, three examples are applied for the empirical analysis of the optimal window length determination method proposed in this study. According to the application results, the proposed method is more applicable than the existing methods. As a result of Kruskal-Wallis rank sum test, it is analyzed that there is no difference in efficiency ranking between DMU(Decision Making Units) according to window length, and the optimal window length selection algorithm can be used for DEA window analysis. In the previous study, it is assumed that the largest number of DMU in the comparison will have a high discriminative power on the efficiency values without a clear discriminant evaluation criterion. In this paper, we propose a method to determine optimal window length using Entropy based discriminant evaluation criterion to solve this problem. According to the result of applying the optimal window length determination method proposed in this study, it is proved that the window length in the case of the largest number of comparison times at one time is the optimal window length.

Statistical and Entropy Based Human Motion Analysis

( Chin-poo Lee ),( Wei-lee Woon ),( Kian-ming Lim ) 한국인터넷정보학회 2010 KSII Transactions on Internet and Information Syst Vol.4 No.6

As visual surveillance systems gain wider usage in a variety of fields, it is important that they are capable of interpreting scenes automatically, also known as “human motion analysis” (HMA). However, existing HMA methods are too domain specific and computationally expensive. This paper proposes a general purpose HMA method that is based on the idea that human beings tend to exhibit erratic motion patterns during abnormal situations. Limb movements are characterized using the statistics of angular and linear displacements. In addition, the method is enhanced via the use of the entropy of the Fourier spectrum to measure the randomness of subject`s motions. Various experiments have been conducted and the results indicate that the proposed method has very high classification accuracy in identifying anomalous behavior.

Entropy와 PCA-DEA 모형을 이용한 은행 대출상담사의 서비스 품질 효율성 분석

최장기,김경택,서재준 한국산업경영시스템학회 2017 한국산업경영시스템학회지 Vol.40 No.3

Loan consultants assist clients with loan application processing and loan decisions. Their duties may include contacting people to ask if they want a loan, meeting with loan applicants and explaining different loan options. We studied the efficiency of service quality of loan consultants contracted to a bank in Korea. They do not work as a team, but do work independently. Since he/she is not an employee of the bank, the consultant is paid solely in proportion to how much he/she sell loans. In this study, a consultant is considered as a decision making unit (DMU) in the DEA (Data Envelopment Analysis) model. We use a principal component analysis-data envelopment analysis (PCA-DEA) model integrated with Shannon’s Entropy to evaluate quality efficiency of the consultants. We adopt a three-stage process to calculate the efficiency of service quality of the consultants. In the first stage, we use PCA to obtain 6 synthetic indicators, including 4 input indicators and 2 output indicators, from survey results in which questionnaire items are constructed on the basis of SERVQUAL model. In the second stage, 3 DEA models allowing negative values are used to calculate the relative efficiency of each DMU. In the third stage, the weight of each result is calculated on the basis of Shannon’s Entropy theory, and then we generate a comprehensive efficiency score using it. An example illustrates the proposed process of evaluating the relative quality efficiency of the loan consultants and how to use the efficiency to improve the service quality of the consultants.

Emergy analysis-based study of the sustainable development of Kunming’s urban eco-economic system

Li Hui,Cao Lei,Chen Youbin,Qiu Huiyi,Huang Xue,Guo Wanqi,Wang Shuntao,Sheng Weiguo,Zhou Yilu,Xu Qing 서울시립대학교 도시과학연구원 2023 도시과학국제저널 Vol.27 No.2

We studied Kunming’s urban eco-economic system’s evolution from 2009 to 2018, using the emergy theory and method; with the entropy evaluation method, we analyzed the system’s natural, social and economic subsystems to know its sustainable development capacity. The results show that the system’s emergy structure has clearly improved, and its renewable resources’ emergy has been rationally employed. The environment load ratio, after reaching its peak in 2011, has been decreasing, and the environment stress has becoming lower year by year, as is conducive to sustainable development. Its eco-economic system, however, is highly dependent on the internal resources; its emergy self-sufficiency ratio indicates that it does not see sufficient utilization of imported emergy and feels excessive stress on its internal resources. It witnesses relatively high emergy intensity. The current population is about twice of its renewable resources’ population carrying capacity, and its dramatically growing population is presenting even high pressure on its social subsystems. This indicates that the living standard per capita in Kunming is lower than the national average. All of these factors will hinder the sustainable development of Kunming’s urban eco-economic system. Our analysis of attribute classification shows that it is more appropriate to calculate the tourist revenue by taking into account export emergy. At the same time, the analysis results of sustainable development emergy indexes obtained by ESI, EISD, euehi, euehi’ and entropy method are compared, which shows that the entropy method is able to better reflect the dynamic changes of the sustainable development capacity of its urban eco-economic system because it introduces all indicators of social, economic and natural subsystems for evaluation. In a word, sustainable development capacity represents the outcome of tradeoff and coordination of the three subsystems of society, economy and nature, and we can formulate corresponding strategies based on the above research findings.

Shannon’s Entropy와 DEA의 결합모형을 이용한 국내 은행의 효율성 분석

김재윤(Jaeyun Kim),심광식(Gwangsic Sim),최성호(Sungho Choi) 한국산업경제학회 2016 산업경제연구 Vol.29 No.2

본 연구는 최근 국내 은행산업의 가장 큰 특징인 대형화와 겸업화가 은행의 효율성에 미치는 영향을 비모수적 접근방법(Non-Parametric Approach)인 자료봉합법(Data Envelopment Analysis: DEA)과 다양한 DEA 모형의 효율성 결과 값들을 모두 고려하여 DMU별 최종 효율성 값을 도출하는 모형인 Shannon’s Entropy의 결합모형을 이용하여 분석하였다. 본 연구에서는 금융중개기능 접근법(Intermediation Approach)에 기초하여 1999년부터 2013년까지 국내 17 은행(7개 시중은행, 6개 지방은행, 4개 특수은행)의 자료를 이용하여 DEA 모형별 효율성을 추정하였고, Shannon’s Entropy를 이용하여 국내 은행의 최종 효율성을 산출하였다. 분석결과, 국민은행과 산업은행, 그리고 농협은 연도별로 고르게 효율성 값이 상위권에 위치하였고, 광주은행과 전북은행, 그리고 제주은행은 효율성이 지속적으로 낮게 나타났다. 시중은행 중에서는 국민은행의 15개년 동안 평균 효율성이 0.997로 1위를 차지하였고, 특수은행 중에서는 농협의 평균 효율성이 0.965로 높게 분석되었고, 지방은행 중에서는 대구은행이 평균 0.294의 효율성 값을 보이며 지방은행 중에서는 가장 높게 나타났다. 분석 결과를 종합해 보면, 상대적으로 규모가 큰 시중은행과 특수은행의 효율성이 다른 지방은행에 비해서 상대적으로 높게 나타났고, 지방은행의 낮은 효율성은 규모면에서 상대적으로 작은 지방은행이 직면하고 있는 규모의 비효율성에서 기인한다고 추측할 수 있다. 따라서 지방은행이 효율적인 자본중개기능을 수행하기 위해서는 은행의 규모를 확장함으로써 규모의 효율성을 제고할 필요가 있고 이를 위한 보다 많은 지방은행 자체의 노력과 함께 금융당국의 정책적 의지가 필요해 보인다. This study investigates the efficiency of 17 Korean banks (7 nationwide commercial banks, 6 provincial banks, and 4 specialized banks) using the Data Envelopment Analysis (DEA), which is a non-parametric method widely used for efficiency assessment of units having similar characteristics. However, the DEA method has a problem of making the choice of the specific model difficult. One way of overcoming this difficulty is Shannon’s Entropy measure and it combines the DEA models in such a way that all the efficiency results calculated from each model are considered in the integrated model to rank the units. Therefore, the paper examines the Korean bank efficiency from 1999 to 2013 by employing the methodology based upon Shannon’s entropy formula for combining the efficiency results of different DEA models (CCR-I, BCC-I, BCC-O, FDH-I, and SBM-I). The empirical results show that Kookmin bank, Korea Development bank, and NongHyup bank are the most efficient banks among Korean banks throughout our sample period, while Kwangju bank, Jeonbuk bank, and Jeju bank are the least efficient banks. Among nationwide commercial banks, the average efficiency of Kookmin bank for 15 years is the highest at 0.997. NongHyup bank has the average efficiency score of 0.965 and it achieves the highest efficiency among specialized banks. Even though all provincial banks have a low efficiency score, Daegu bank has 0.294 which is the highest level among provincial banks. Overall, while all nationwide commercial banks and specialized banks which are relatively big in size have high efficiency scores, all provincial banks turn out to be inefficient. We conjecture that the low level of efficiency of provincial banks may be caused by the diseconomy of scale. Therefore, it is very important that to improve the efficiency of provincial banks, it is necessary to increase the size of provincial banks through their own efforts. In addition, the Financial Services Authority needs to take an appropriate action to support provincial banks.