• Journal of the Korean Data and Information Science Society
• /
• v.20 no.3
• /
• pp.515-526
• /
• 2009

Pricing based on proper risk has been one of main issues in auto insurance. In this paper, we review how the techniques of pricing in auto insurance have been developed and suggest a better approach which meets the existing risk statistically by comparison. The generalized linear model (GLM) method is discussed for pricing with different distributions. With GLM approach, the distribution of error assumed plays an main role for the best fit corresponding to the characteristics of dependent variables. Tweedie distribution is considered as one of error distributions in addition to widely used Gamma and Poisson distribution. With these different types of error assumption for estimating the proper premium in auto insurance, various modeling approaches are possible. In this paper, various modeling approaches with different assumptions for estimating proper risk is discussed and also real example is given by assuming different.

• Journal of the Korean Society of Safety
• /
• v.37 no.6
• /
• pp.89-95
• /
• 2022

A method for measuring the length and slope of a temporary structure using an unmanned aerial vehicle (UAV) and 3D modeling method is proposed. The actual length and slope of the vertical member of the specimen were measured and compared with the measured values obtained by the proposed method for the specimens with and without the vertical protection net installed. Based on the result of measuring the length of the temporary structure specimen using the UAV and 3D modeling method, the measured value showed an error of 0.87% when compared to the actual length in the specimen without the vertical protection net installed. In addition, the error of the slope was 0.63°. It was thought that the proposed method could be usable for the purpose of finding parts in wrong installation state on the temporary structure and informing the manager in charge. However, in the case of the specimen with the vertical protection net, the measurement showed a 1.46% error in length and 2.77° difference in slope. Therefore, if a vertical protection net is to be installed in a temporary structure, the measurement accuracy should be improved by utilizing an image processing method, etc.

• Journal of radiological science and technology
• /
• v.46 no.2
• /
• pp.131-139
• /
• 2023

High-dose I-131 used for the treatment of thyroid cancer causes localized exposure among radiology technologists handling it. There is a delay between the calibration date and when the dose of I-131 is administered to a patient. Therefore, it is necessary to directly measure the radioactivity of the administered dose using a dose calibrator. In this study, we attempted to apply machine learning modeling to measured external dose rates from shielded I-131 in order to predict their radioactivity. External dose rates were measured at 1 m, 0.3 m, and 0.1 m distances from a shielded container with the I-131, with a total of 868 sets of measurements taken. For the modeling process, we utilized the hold-out method to partition the data with a 7:3 ratio (609 for the training set:259 for the test set). For the machine learning algorithms, we chose linear regression, decision tree, random forest and XGBoost. To evaluate the models, we calculated root mean square error (RMSE), mean square error (MSE), and mean absolute error (MAE) to evaluate accuracy and R² to evaluate explanatory power. Evaluation results are as follows. Linear regression (RMSE 268.15, MSE 71901.87, MAE 231.68, R² 0.92), decision tree (RMSE 108.89, MSE 11856.92, MAE 19.24, R² 0.99), random forest (RMSE 8.89, MSE 79.10, MAE 6.55, R² 0.99), XGBoost (RMSE 10.21, MSE 104.22, MAE 7.68, R² 0.99). The random forest model achieved the highest predictive ability. Improving the model's performance in the future is expected to contribute to lowering exposure among radiology technologists.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.23 no.1
• /
• pp.167-181
• /
• 2024

Utilization of a driving simulator in the development of autonomous driving technology allows us to perform various tests effectively in criticial environments, thereby reducing the development cost and efforts. However, there exists a serious drawback that the driving simulator has a big difference from the real environment, so a problem occurs when the autonomous driving algorithm developed using the driving simulator is applied directly to the real vehicle system. This is defined as so-called Sim2Real problem and can be classified into scenarios, sensor modeling, and vehicle dynamics. This Paper presensts on a method to solve the Sim2Real problem in autonomous driving simulator focusing on IMU sensor. In order to reduce the difference between emulated virtual IMU sensor real IMU sensor, IMU sensor emulation techniques through precision error modeling of IMU sensor are introduced. The error model of IMU sensors takes into account bias, scale factor, misalignmnet, and random walk by IMU sensor grades.

• Journal of Electrical Engineering and Technology
• /
• v.7 no.5
• /
• pp.653-658
• /
• 2012

Load modeling has a significant influence on power system analysis and control. In recent years, measurement-based load modeling has been widely practiced. In the load modeling algorithm, the model structure is determined and the parameters of the established model are estimated. For parameter estimation, least-squares optimization method is applied. The model parameters are estimated so that the error between the measured values and the predicted values is to be minimized. By introducing sliding window concept, on-line load modeling method can be performed which reflects the dynamic behaviors of loads in real-time. For the purpose of data acquisition, the measurement system including PMU is implemented in university level. In this paper, case studies are performed using real PMU data from Korea Univ. and Seoul National University of Science and Technology. The performances of modeling real and reactive power behaviors using exponential and ZIP load model are evaluated.

• 제어로봇시스템학회:학술대회논문집
• /
• 2002.10a
• /
• pp.102.2-102
• /
• 2002

In this paper, the improved Multi-FNN (Fuzzy-Neural Networks) model is identified and optimized using HCM (Hard C-Means) clustering method and optimization algorithms. The proposed Multi-FNN is based on FNN and use simplified and linear inference as fuzzy inference method and error back propagation algorithm as learning rules. We use a HCM clustering and genetic algorithms (GAs) to identify both the structure and the parameters of a Multi-FNN model. Here, HCM clustering method, which is carried out for the process data preprocessing of system modeling, is utilized to determine the structure of Multi-FNN according to the divisions of input-output space using I/O process data. Also, the parame...

• 제어로봇시스템학회:학술대회논문집
• /
• 1991.10a
• /
• pp.905-910
• /
• 1991

This paper presets a leakage detection method based on modeling the leakage in pipeline systems. For gas pipeline systems, a method based on the state space model is suggested. For liquid pipeline systems, an experiment based on the static model equation was performed. In the experiment, it was possible to detect the leak and to diagnosis the leak situation within the error of .+-.3%.

• Proceedings of the KIPE Conference
• /
• 2003.07b
• /
• pp.913-916
• /
• 2003

In many cases, ac line EMI filters are designed by trial and error method. Thus, designing a filter is a time-consuming process. To overcome the problem, the newly analytical design procedure is proposed in this paper. Using the modeling circuit of each filter components, the filter design is carried out applicable to both common-mode (CM) and differential-mode (DM). These are verified through the experimental results.

• 제어로봇시스템학회:학술대회논문집
• /
• 2000.10a
• /
• pp.39-39
• /
• 2000

In this paper, a decoupler of tandem cold rolling mill is designed. Before designing the decoupler, this paper improved conventional linear model by considering friction and yield stress of rolling strip. In a stand, the decoupler let an output be controlled by an input. And even if states of other stands should be changed, current stand takes no interference from those changes. In addition, with the same method, a feedforward controller is designed for an input strip thickness error. Finally, performance of controllers above is shown with nonlinear simulation.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.49 no.11
• /
• pp.625-631
• /
• 2000

In designing fuzzy model, we encounter a major difficulty in the identification of an optimized fuzzy rule base, which is traditionally achieved by a tedious-and-error process. This paper presents an approach to automatic design of optimal fuzzy rule bases for modeling using evolutionary programming. Evolutionary programming evolves simultaneously the structure and the parameter of fuzzy rule base a given task. To check the effectiveness of the suggested approach, 3 examples for modeling are examined, and the performance of the identified models are demonstrated.