• 산업공학
• /
• 제20권4호
• /
• pp.448-457
• /
• 2007

The small & micro business has the characteristics of both consumer credit risk and business credit risk. In predicting the bankruptcy for small-micro businesses, the problem is that in most cases, the financial data for evaluating business credit risks of small & micro businesses are not available. To alleviate such problem, we propose a bankruptcy prediction mechanism using the credit card sales information available, because most small businesses are member store of some credit card issuers, which is the main purpose of this study. In order to perform this study, we derive some variables and analyze the relationship between good and bad signs. We employ the new statistical learning technique, support vector machines (SVM) as a classifier. We use grid search technique to find out better parameter for SVM. The experimental result shows that credit card sales information could be a good substitute for the financial data for evaluating business credit risk in predicting the bankruptcy for small-micro businesses. In addition, we also find out that SVM performs best, when compared with other classifiers such as neural networks, CART, C5.0 multivariate discriminant analysis (MDA), and logistic regression.

• 동력기계공학회지
• /
• 제6권1호
• /
• pp.55-60
• /
• 2002

There has been no report on the life prediction for gas turbine materials at high temperatures based on the creep properties and their relationship with the AE(acoustic emission) properties as a means of real-time non-destructive testing. One of the important issues is thus to develop a reliable method of evaluating creep properties by the AE technique. In this paper, the real-time evaluation of high temperature creep time and AE cumulative counts for nickel-based superalloy Udimet 720 was performed on round-bar type specimens under pure load at the temperatures of 811, 922 and 977K. The total AE cumulative counts until the starting point of secondary creep($N_1$) and that of tertiary creep($N_2$) have quantitative relationship with the tertiary creep time and the rupture time. It is thus possible to construct the life prediction system based on creep and the prevention system of tertiary creep by using AE technique.

• 한국기계가공학회지
• /
• 제20권9호
• /
• pp.90-96
• /
• 2021

As real-time data of factories can be collected using various sensors, the adaptation of intelligent unmanned processing systems is spreading via the establishment of smart factories. In intelligent unmanned processing systems, data are collected in real time using sensors. The equipment is controlled by predicting future situations using the collected data. Particularly, a technology for the prediction of tool wear and for determining the exact timing of tool replacement is needed to prevent defected or unprocessed products due to tool breakage or tool wear. Directly measuring the tool wear in real time is difficult during the cutting process in milling. Therefore, tool wear should be predicted indirectly by analyzing the cutting load of the main spindle, current, vibration, noise, etc. In this study, data from the current and acceleration sensors; displacement data along the X, Y, and Z axes; tool wear value, and shape change data observed using Newroview were collected from the high-speed, two-edge, flat-end mill machining process of SKD11 steel. The support vector machine technique (machine learning technique) was applied to predict the amount of tool wear using the aforementioned data. Additionally, the prediction accuracies of all kernels were compared.

• International Journal of Computer Science & Network Security
• /
• 제22권3호
• /
• pp.29-36
• /
• 2022

In deep learning classification tasks, most models frequently assume that all labels are available for the training datasets. As such strategies to learn new concepts from unlabeled datasets are scarce. In fingerprint classification tasks, most of the fingerprint datasets are labelled using the subject/individual and fingerprint datasets labelled with finger type classes are scarce. In this paper, authors have developed approaches of classifying fingerprint images using the majorly known fingerprint classes. Our study provides a flexible method to learn new classes of fingerprints. Our classifier model combines both the clustering technique and use of deep learning to cluster and hence label the fingerprint images into appropriate classes. The K means clustering strategy explores the label uncertainty and high-density regions from unlabeled data to be clustered. Using similarity index, five clusters are created. Deep learning is then used to train a model using a publicly known fingerprint dataset with known finger class types. A prediction technique is then employed to predict the classes of the clusters from the trained model. Our proposed model is better and has less computational costs in learning new classes and hence significantly saving on labelling costs of fingerprint images.

• Structural Engineering and Mechanics
• /
• 제81권1호
• /
• pp.103-115
• /
• 2022

The prediction error variances for frequencies are usually considered as unknown in the Bayesian system identification process. However, the error variances for mode shapes are taken as known to reduce the dimension of an identification problem. The present study attempts to explore the effectiveness of Bayesian approach of model parameters updating using Markov Chain Monte Carlo (MCMC) technique considering the prediction error variances for both the frequencies and mode shapes. To remove the ergodicity of Markov Chain, the posterior distribution is obtained by Gaussian Random walk over the proposal distribution. The prior distributions of prediction error variances of modal evidences are implemented through inverse gamma distribution to assess the effectiveness of estimation of posterior values of model parameters. The issue of incomplete data that makes the problem ill-conditioned and the associated singularity problem is prudently dealt in by adopting a regularization technique. The proposed approach is demonstrated numerically by considering an eight-storey frame model with both complete and incomplete modal data sets. Further, to study the effectiveness of the proposed approach, a comparative study with regard to accuracy and computational efficacy of the proposed approach is made with the Sequential Monte Carlo approach of model parameter updating.

• 한국축산식품학회지
• /
• 제38권5호
• /
• pp.1109-1119
• /
• 2018

In this paper, we report the development of a nondestructive prediction model for lean meat percentage (LMP) in Korean pig carcasses and in the major cuts using a machine vision technique. A popular vision system in the meat industry, the VCS2000 was installed in a modern Korean slaughterhouse, and the images of half carcasses were captured using three cameras from 175 selected pork carcasses (86 castrated males and 89 females). The imaged carcasses were divided into calibration (n=135) and validation (n=39) sets and a multilinear regression (MLR) analysis was utilized to develop the prediction equation from the calibration set. The efficiency of the prediction equation was then evaluated by an independent validation set. We found that the prediction equation - developed to estimate LMP in whole carcasses based on six variables - was characterized by a coefficient of determination ($R^2_v$) value of 0.77 (root-mean square error [RMSEV] of 2.12%). In addition, the predicted LMP values for the major cuts: ham, belly, and shoulder exhibited $R^2_v$ values${\geq}0.8$ (0.73 for loin parts) with low RMSEV values. However, lower accuracy ($R^2_v=0.67$) was achieved for tenderloin cuts. These results indicate that the LMP in Korean pig carcasses and major cuts can be predicted successfully using the VCS2000-based prediction equation developed here. The ultimate advantages of this technique are compatibility and speed, as the VCS2000 imaging system can be installed in any slaughterhouse with minor modifications to facilitate the on-line and real-time prediction of LMP in pig carcasses.

• 지질공학
• /
• 제18권1호
• /
• pp.45-54
• /
• 2008

본 연구에서는 기 조사된 편마암 지역에서의 급경사지재해 발생지역 및 미발생지역에 대한 현장조사자료 및 토질시험자료를 토대로 통계적인 분석방법인 의사결정나무모형을 이용하여 급경사지재해 예측기법을 개발하였다. 편마암 지역에서의 조사된 급경사지재해 자료는 서울 및 경기지역에서 1998년 집중호우로 발생된 104개소구간이다. 이 가운데 예측모델 개발에 활용된 자료수는 결측치를 제외한 61개소로서, 급경사지재해 발생구간 34개소와 미발생구간 27개소이다. 의사결정나무모형을 이용한 통계적인 분석은 카이제곱 통계량, 지니 지수 및 엔트로피 지수를 적용하여 실시하였다. 분석결과 사면경사, 포화도 및 사면고도가 분리기준으로 선택되었으며, 엔트로피 지수를 이용한 의사결정나무모형 예측모델이 정확도가 가장 높은 것으로 나타났다. 선정된 급경사지재해 예측모델의 분리기준은 최상위부터 사면경사, 포화도 및 사면고도의 순서로 선택되었으며, 각각의 분리기준치는 사면경사의 경우 $17.9^{\circ}$, 포화도의 경우 52.1%, 사면고도의 경우 320m로 결정되었다.

• 한국정보과학회논문지:데이타베이스
• /
• 제34권6호
• /
• pp.483-492
• /
• 2007

최근 센서 네트워크의 발달로 실세계의 많은 데이타가 시간 속성을 갖고 실시간으로 수집되고 있다. 기존의 시계열 데이타 예측 기법은 모델 갱신 없이 예측을 수행하였다. 그러나 스트림 데이타는 매우 빠르게 수집이 되고 시간이 지남에 따라 데이타의 특성이 변경될 수 있으므로 기존의 시계열 예측 기법을 적용하는 것은 적절하지 않다. 따라서 이 논문에서는 슬라이딩 윈도우와 점진적인 회귀분석을 이용한 스트림 데이타 예측 기법을 제안한다. 이 기법은 스트림 데이타를 다중 회귀 모델에 입력하기 위해 차원 분열을 통해 여러 개의 속성으로 분열(Fractal)하고, 변화되는 데이타의 분포를 반영하기 위해 슬라이딩 윈도우 기법을 사용하여 점진적으로 회귀 모델을 갱신한다. 또한 고정 크기 큐를 이용하여 최근의 데이타로만 모델을 유지한다. 이전 데이타의 유지 없이 최소 정보를 갖는 행렬을 통해 모델을 갱신하므로 낮은 공간 복잡도를 갖고 점진적으로 모델을 갱신함으로써 에러율의 증가를 방지한다. 제안된 기법의 타당성은 RME(Relative Mean Error)와 RMSE(Root Mean Square Error)를 이용하여 측정하였고, 실험 결과 다른 기법에 비해 우수하였다.

• KIEE International Transactions on Power Engineering
• /
• 제11A권4호
• /
• pp.15-20
• /
• 2001

Power system stability is correlated with system structure, disturbances and operating conditions, and power flows on transmission lines are closely related with those conditions. This paper proposes a methodology to identify correlative power flows for power system transient and small-signal stability prediction. In transient stability sense, the Critical Clearing Time is used to select some dominant contingencies, and Transient Stability Prediction index is proposed for the quantitative comparison. For small-signal stability discusses a methodology to identify crucial transmission lines for stability prediction by introducing a sensitivity factor based on eigenvalue sensitivity technique. On-line monitoring of the selected lines enables to predict system stability in real-time. Also, a procedure to make a priority list of monitored transmission lines is proposed. The procedure is applied to a test system, and it shows capabilities of the proposed method.

• 전기학회논문지
• /
• 제62권11호
• /
• pp.1590-1597
• /
• 2013

Electric power is an important part in economic development. Moreover, an accurate load forecast can make a financing planning, power supply strategy and market research planned effectively. This paper used the fuzzy logic system to predict the regional electric power load. To design the fuzzy prediction system, the correlation-based clustering algorithm and TSK fuzzy model were used. Also, to improve the prediction system's capability, the moving average technique and relative increasing rate were used in the preprocessing procedure. Finally, using four regional electric power load in Taiwan, this paper verified the performance of the proposed system and demonstrated its effectiveness and usefulness.