Exploring class imbalance with under-sampling, over-sampling, and hybrid sampling based on Mahalanobis distance for landslide susceptibility assessment: a case study of the 2018 Iburi earthquake induced landslides in Hokkaido, Japan

남경훈,김종태,채병곤 한국지질과학협의회 2024 Geosciences Journal Vol.28 No.1

This study focuses on evaluating the performance of the resampling approach using under-sampling, over-sampling, and hybrid sampling techniques in the random forest (RF) model for landslide susceptibility assessment (LSA). For this research, the study area selected was Hokkaido, Japan, which experienced a total of 5,625 landslides as a single event caused by the 2018 Ibury earthquake. The objective of this study is to address the class imbalance issue and improve the accuracy of LSA. Multiple data sources are utilized to obtain conditioning factors, and objective absence data sampling based on Mahalanobis distance is employed to tackle the unlabeled sample problem. The RF model is used to calculate landslide susceptibility values and generate LSA. These values are then evaluated using two diagnostic tools, the Area Under the Receiver Operating Characteristic curve (AUROC) and the Precision-Recall curve (AUPRC). These tools help validate and interpret binary classification predictive models for imbalanced data. The results demonstrate improved performance with larger sample sizes, and the resampling approach yields better consistency compared to random sampling within the study area. To enhance the accuracy and consistency of machine learning techniques in reducing landslide risks, the study recommends utilizing hybrid sampling technique and Mahalanobis distance-based absence data sampling in LSA.

Cook-type Distance in Multivariate Constrained Regression

김명근 한국자료분석학회 2009 Journal of the Korean Data Analysis Society Vol.11 No.2

An influence measure, which we will call Cook-type distance, for multivariate regression with linear constraints on regression coefficients is suggested. It is a modification of the usual Cook's distance adapted to multivariate constrained regression and it is defined by a distance between the estimates of regression coefficients evaluated with and without some observations, scaled by the covariance matrix of the distance, not by that of the estimates of regression coefficients as in the usual Cook's distance. It would be used for detecting influential observations. Its sampling distribution will be derived, while that of the usual Cook's distance is not available. Thus we can judge whether each observation is influential or not for multivariate constrained regression, while we can not use the cutoff values, based on the usual Cook's distance, so that the use of the usual Cook's distance is a rule of thumb. A numerical example is given for illustration.

가변적인 샘플링을 이용한 차원 감소법에 의한 신뢰도 해석 기법

육순민(Sunmin Yook),민준홍(Junhong Min),김동호(Dong-Ho Kim),최동훈(Dong-Hoon Choi) 대한기계학회 2009 大韓機械學會論文集A Vol.33 No.9

This study provides how the Dimension Reduction (DR) method as an efficient technique for reliability analysis can acquire its increased efficiency when it is applied to highly nonlinear problems. In the highly nonlinear engineering systems, 4N+1 (N: number of random variables) sampling is generally recognized to be appropriate. However, there exists uncertainty concerning the standard for judgment of non-linearity of the system as well as possibility of diverse degrees of non-linearity according to each of the random variables. In this regard, this study judged the linearity individually on each random variable after 2N+1 sampling. If high non-linearity appeared, 2 additional sampling was administered on each random variable to apply the DR method. The applications of the proposed sampling to the examples produced the constant results with increased efficiency.

가변적인 샘플링을 이용한 신뢰도 해석 기법

육순민(Sunmin Yook),최동훈(Dong-Hoon Choi) 대한기계학회 2008 대한기계학회 춘추학술대회 Vol.2008 No.11

This study provides how the Dimension Reduction (DR) method as an efficient technique for reliability analysis can acquire its increased efficiency when it is applied to highly nonlinear problems. In the highly nonlinear engineering systems, 4N+1 (N: number of random variables) sampling is generally recognized to be appropriate. However, there exists uncertainty concerning the standard for judgment of non-linearity of the system as well as possibility of diverse degrees of non-linearity according to each of the random variables. In this regard, this study judged the linearity individually on each random variable after 2N+1 sampling. If high non-linearity appeared, 2 additional sampling was administered on each random variable to apply the DR method. The applications of the proposed sampling to the examples produced the constant results with increased efficiency.

ACTIVE ROLL PREVIEW CONTROL WITH V2V COMMUNICATION

Seongjin Yim 한국자동차공학회 2019 International journal of automotive technology Vol.20 No.1

This paper presents an active roll preview control with vehicle-to-vehicle (V2V) communication. In intelligent transportation system (ITS) environments, several states and measured signals can be transmitted to following rear vehicle through V2V communication device. The preceding vehicle’s lateral acceleration is sent to the following vehicle. A LQ optimal preview controller for application to active roll control(ARC) is designed using the lateral acceleration. To cope with the speed change of vehicles, a position-based sampling on lateral acceleration is adopted. Based on the data, a distance-based interpolation is performed to resample the time-based lateral acceleration for preview control of ARC. To check the performance of the proposed method, simulation is performed with linear and nonlinear vehicle models.

경기도를 방문한 외래 관광객들의 관광활동 참여 패턴 비교 - 시간적 거리 관점에서

박득희(Park, Deuk-Hee),강상훈(Kang, Sang-Hoon) 경기연구원 2021 GRI 연구논총 Vol.23 No.3

This study compared participation patterns of tourism activities by international tourists who visited Gyeonggi province in South Korea from a temporal distance perspective using social network analysis techniques. The raw data of 2019 Gyeonggi province international tourist survey was applied. Of diverse social network analysis techniques, this study performed density analysis, normalized eigenvector centrality, boostrap paried sample t-test, and link analysis. Analysis results are: first, based on the density analysis result, there were differences in network structure between two groups, such as determination of the visit to Gyeonggi province based on temporal distance; second, normalized eigenvector centrality scores indicate differences in the ranks among tourism activities between the two groups. These results demonstrate that differences in tourism activities and the role of tourism activities by international tourists who visited Gyeonggi province between make a decision before visit Korea and after visit Korea. Lastly, link analysis results show differences in participation patterns of tourism activities according to international tourists’ temporal distance. Through these findings, this study provide theoretical implications associated with tourism activity participation patterns as well as practical implications for the sustainable tourism growth of Gyeonggi province, Future research suggestions also provided in the conclusion.

THE CALIBRATION ESTIMATION USING TWO-STEP NEWTON'S ALGORITHM IN TWO-PHASE SAMPLING

Son, Chang-Kyoon,Yum, Joon-Keun 한국전산응용수학회 2000 Journal of applied mathematics & informatics Vol.7 No.1

In this paper, we consider to the adjustment weighting procedure in the two phase sampling scheme. In general, the unit nonresponses may be occured in the final survey operation. When the unit nonresponse be generated in survey, it is able to use the auxiliary variable for estimating of interest variable. In this viewpoint, we use the two kinds level of auxiliary variable, $X_{1k}$ and $X_{2k}$ for the calibration procedure. We proprose the two-step Newton's method in the calibration estimation procedure for the two phase sampling.

브레이크 로봇을 활용한 제동안정성평가 신뢰성 향상방안 연구

조성우(Seongwoo Cho),최인성(Inseong Choi),안장모(Jangmo Ahn),김성섭(Seongsub Kim) 한국자동차공학회 2010 한국자동차공학회 부문종합 학술대회 Vol.2010 No.5

The Korean New Car Assessment Program (KNCAP) has been carried out by the Korean government since 1999 in order to encourage that the auto makers shall launch the safer cars into the domestic market. The brake NCAP test has been adopted to reduce the stopping distance. Braking tests were performed with pedal force applied by driver/robot individually and their effects upon pedal force requirements were analyzed. The differences of stopping distance were significantly shorter with pedal force applied by brake robot compared to pedal force applied by driver. Accordingly, this study evaluated the effectiveness of brake robot that will reduce the deviation of stopping distances through pedal force requirements of brake NCAP. The methodology to integrate the respective ratings from the braking tests considering environmental variables would be studied for the future application.

Testing the potential of lightweight drones as a tool for monitoring the status of colonially breeding Saunders's gulls (Saundersilarus saundersi)

Han-I Choi(최한이),Hyung-Kyu Nam(남형규),Jongmin Yoon(윤종민) 한국조류학회 2020 한국조류학회지 Vol.27 No.1

무인항공기는 최근 물새류의 번식지를 포함한 다양한 야생 동물 현황을 조사하는데 사용되고 있으며, 비침습적인 방식으로 야생동물의 개체수와 분포에 대한 조사 데이터를 얻는데 용이하다. 검은머리갈매기(Saundersilarus saundersi)는 갯벌 주변 매립지에서 집단 번식하는 물새류로, 집단 방어행동으로 인해 둥지를 직접 조사 등 인위적인 방해 없이 번식현황을 파악하기 어려운 것으로 알려져 있다. 본 연구에서는 인천 매립지 서식 검은머리갈매기 집단 번식지에서 소형 드론을 사용하여 번식 현황을 조사하였다. 2018년, 선조사 방식의 Distance Sampling 방법으로 번식기 동안 반복 조사를 실시했다. 드론으로 녹화된 비디오 데이터는 조사 선에서 개체 또는 둥지까지의 거리로 변환하여 기록하였다. 결과적으로, 번식기 동안 집단 번식지의 밀도가 급격히 감소하였으며, 2018년 집단 번식은 둥지포식에 의해 완전한 실패하는 패턴을 보였다. 따라서, Distance Sampling 기반으로 비침습적이고 저비용의 무인항공기 조사방법은 데이터 수집과 추정에 불확실성이 존재하지만, 물새류 집단 번식지 모니터링에 유용하다고 추론된다. Unmanned aerial vehicles (UAV) are recently used to survey a diversity of wildlife status, including breeding colonies of seabirds, which provides empirical data of wildlife numbers and distributions with fewer disturbances. The Saunders’s gull (Saundersilarus saundersi) is a colonially breeding seabird species on reclaimed lands near mudflats. It is also known to be difficult to estimate the breeding status without disturbance such as directly checking nests due to their group defense behavior. In this study, we used a lightweight drone to survey the breeding status of Saunders’s gulls in a reclaimed area of Incheon, Republic of Korea. Then, we approached a distance sampling method with fixed line transects along multiple surveys during the breeding season of 2018. The video data recorded by UAV were transcribed into individual- or nest-based distances from each line transect. The results of our survey indicated that the relative size of breeding colony rapidly decreased throughout the breeding season, showing the complete failure pattern of the breeding colony in 2018. We suggest that this non-invasive, less expensive UAV method with distance sampling may be useful to estimate the breeding status of seabird colonies in spite of uncertainty of data collection and estimation.

An Estimation of Long-term Settlements in the Large Reclamation Site and Determination of Additional Sampling Positions Using Geostatistics and GIS

이혁진(Lee Hyeog Jin),박사원(Park Sa Won),유시동(Yu Si Dong),김홍택(Kim Hong Taeg) 한국지반공학회 2004 한국지반공학회논문집 Vol.20 No.2

For geotechnical applications, engineers use data obtained from a site investigation to interpret the structure and potential behavior of the subsurface. In most cases, these data consist of samples that represent 1/100,000 or less of the total volume of soil. These samples and associated field and lab testing provide the information used to estimate soil parameter values. The resulting values are estimated ones and there exists some likelihood that actual soil conditions are significantly different from the estimates. This may be the case even if the sampling and interpretation procedures are performed in accordance with standard practice. Although these efforts have been made to characterize the uncertainty associated with geotechnical parameters, there is no commonly accepted method to evaluate quantitatively the quality of an investigation plan as a whole or the relative significance of individual sampling points or potential sampling points. In this paper, the geostatistical methods, Inverse distance weighted interpolation method and kriging interpolation method for estimating soil parameter values, are compared with each other to examine the accuracy of these methods. And new method for estimating the long-term settlements is proposed. This method combines the concept of investigation plan thoroughness with geostatistical techniques to develop a spatially sensitive estimate of the site investigation plan quality, referred to as continuous thoroughness, because it applies to geotechnical parameters that exist at every point throughout the sampling area. Finally, the method for locating additional sampling positions is proposed. The approach was implemented using a geographic information system platform(ARC-Info) so that the visual, analytical, and data management benefits of GIS could be utilized for a more detailed analysis.