• The Korean Journal of Applied Statistics
• /
• v.31 no.2
• /
• pp.229-239
• /
• 2018

This paper introduces how to adopt the concept of temporal hierarchies to forecast time series data. Similarly as in hierarchical cross-sectional data, temporal hierarchies can be constructed for any time series data by means of non-overlapping temporal aggregation. Reconciliation forecasts with temporal hierarchies result in more accurate and robust forecasts when compared with the independent base and bottom-up forecasts. As an empirical example, we forecast traffic accident counts with temporal hierarchies and observe that reconciliation forecasts are superior to the base and bottom-up forecasts in terms of forecast accuracy.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.35 no.5
• /
• pp.1135-1146
• /
• 2015

This thesis aims to improve the safety of bicycle traffic for activating the use of bicycle, main means of non-powered and non-carbon transportation in order to cope with worldwide crisis such as climate change and energy depletion and to implement sustainable traffic system. In this regard, I analyzed the problem of bicycle roads currently installed and operated, and developed the bicycle accident forecasting model. Following are the processes for this. First, this study presented the current status of bicycle road in Korea as well as accident data, collect the data on bicycle traffic accidents generated throughout the country for recent 3 years (2009~2011) and analyzed the features of bicycle traffic accidents based on the data. Second, this study selected the variable affecting the number of bicycle accidents through accident feature analysis of bicycle accidents at Jeollanam-do, and developed accident forecast model using the multiple regression analysis of 'SPSS Statistics 21'. At this time, the number of accidents due to extension per road types (crossing, crosswalk, other single road) was used. To verify the accident forecast model deduced, this study used the data on bicycle accident generated in Gwangju, 2011, and compared the prediction value with actual number of accidents. As a result, it was found out that reliability of accident forecast model was secured through reconciling with actual number of cases except certain data. Third, this study carried out field survey on the bicycle road as well as questionnaire on satisfaction of bicycle road and use of bicycle for analysis of bicycle road problems, and presented safety improvement measures for the problems deduced as well as bicycle activation plans. This study is considered to serve as the fundamental data for planning and reorganizing of bicycle road in the future, and expected to improve safety of bicycle users and to promote activation of bicycle use as the means of transportation.

• Journal of the Korean Society of Systems Engineering
• /
• v.18 no.2
• /
• pp.75-93
• /
• 2022

Machine learning (ML) data-driven meta-model is proposed as a surrogate model to reduce the excessive computational cost of the physics-based model and facilitate the real-time prediction of a nuclear power plant's transient response. To forecast the transient response three machine learning (ML) meta-models based on recurrent neural networks (RNNs); specifically, Long Short Term Memory (LSTM), Gated Recurrent Unit (GRU), and a sequence combination of Convolutional Neural Network (CNN) and LSTM are developed. The chosen accident scenario is a control element assembly withdrawal at power concurrent with the Loss Of Offsite Power (LOOP). The transient response was obtained using the best estimate thermal hydraulics code, MARS-KS, and cross-validated against the Design and control document (DCD). DAKOTA software is loosely coupled with MARS-KS code via a python interface to perform the Best Estimate Plus Uncertainty Quantification (BEPU) analysis and generate a time series database of the system response to train, test and validate the ML meta-models. Key uncertain parameters identified as required by the CASU methodology were propagated using the non-parametric Monte-Carlo (MC) random propagation and Latin Hypercube Sampling technique until a statistically significant database (181 samples) as required by Wilk's fifth order is achieved with 95% probability and 95% confidence level. The three ML RNN models were built and optimized with the help of the Talos tool and demonstrated excellent performance in forecasting the most probable NPP transient response. This research was guided by the Systems Engineering (SE) approach for the systematic and efficient planning and execution of the research.

• Proceedings of the KOR-KST Conference
• /
• 2007.11a
• /
• pp.315-322
• /
• 2007

신호교차로 교통사고는 90년대 이후 도시가 발달하고 산업이 고도화됨에 따라 교통 혼잡 문제와 함께 심각한 사회문제로 대두되고 있다. 특히 신호교차로의 교통사고는 인적요인, 차량요인, 환경적 요인 등이 복합적으로 작용하여 발생하는데, 교통량의 집중과 도로의 기하구조, 운전자 과실 등이 교통사고의 주요 인자로 작용하고 있다. 본 연구에서 교통사고 예측모형을 개발하기 위해서 2003년부터 2006년도까지 실제 경기도의 신호교차로에서 발생한 교통사고자료를 기초로 하였다. 구체적으로는 시내가 아닌 지방부 성격을 지닌 일반국도를 대상으로 하였다. 지방부 일반국도의 신호교차로 교통사고 분석에 단순통계분석과 다중회귀분석을 사용하였다. 사고와 관계가 높은 신호주기, 방향별 접근 교통량, 회전교통량 둥과 같은 도로, 교통, 운영조건들로 변수를 정하여 교통사고 예측모형을 도출하였다. 본 연구에서는 도로조건, 교통조건, 운영조건들과 사고와의 관계를 이용하여 경기도 일반국도의 신호교차로 교통사고예측모형을 개발하였고, 이는 지방부 성격을 지닌 교차로에 적용이 가능하다고 판단된다.

• Journal of Korean Society for Geospatial Information Science
• /
• v.6 no.1 s.11
• /
• pp.35-51
• /
• 1998

This article aims at presenting a new framework for traffic accident analysis by proposing a new methodology for the management of the accident data and by establishing the relationship between accidents and roadway characteristics within it For the first issue, authors introduced geographic information system (GIS) into the analysis framework of the accident data since it is believed that analysis based on GIS seems to provide more effective information in reducing accidents. Point-based, line-based, and polygon (grid)-based approaches were set of along with concrete examples. Especially, the location-based scores such as localization, specialization coefficients, and Tress score have been added to identify the intensity of certain accident types within study area or grids. The second issue addressed the equation formulation of accident and fatality numbers with roadway characteristics like number of intersections and road length in a grid with a sense that (1) accidents on roadways are the function of the roadway physical characteristics rather than the socio-economical secondary data (2) the equation can be applied to the any 'suggested' area, not just region or nation, and (3) the accident forecasting model should emphasize the accident location itself more than any other factors. Some equations based on those assumption have been derived along with some future research agenda.

• Journal of the Korean Society of Safety
• /
• v.25 no.2
• /
• pp.84-88
• /
• 2010

The main point of this study is to find ways to prevent accidents at complex linear sections in advance by improving geometric structure elements that can be considered from the designing stage. Complex linear roads are consisted of sections where straight sections connect with curved sections or sections where curved sections connect with curved sections with relatively high possibility of accidents and accidents can be reduced through improving designing elements in these sections. Therefore, this study aims to develop accident forecasting model in complex linear roads and to clarify major elements affecting traffic accidents. The results of analysis showed that the groups are divided into a group less than 355m based on curve radius of 355m, a group whose curve radius exceeds 355m and a group whose incline exceeds -0.79 and a group whose curve radius is below 355m and incline exceeds -0.79 for straight section + curved section, and for curved section + curved section, it is divided into a group whose first curved section is less than 410m based on curve radius of 410m and the first curve is turning right and a group exceeding 410m and the first curve is turning left. The major variables common in 2 models are front curve radius and curve types(left, right), road surfaces, weather.

• Proceedings of the KAIS Fall Conference
• /
• 2008.05a
• /
• pp.248-250
• /
• 2008

This paper proposes a prediction method to prevent traffic accident and reduce to vehicle waiting time using neural network. Computer simulation results proved reducing average vehicle waiting time which proposed coordinating green time better than electro-sensitive traffic light system dose not consider coordinating green time. Moreover, we present neural network approach for traffic accident prediction with unnormalized (actual or original collected) data. This approach is not consider the maximum value of data and possible use the network without normalizing but the predictive accuracy is better. Also, the unnormalized method shows better predictive accuracy than the normalized method given by maximum value. Therefore, we can make the best use of this model in software reliability prediction using unnormalized data. Computer simulation results proved reducing traffic accident waiting time which proposed neural network better than conventional system dosen't consider neural network.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.30 no.4D
• /
• pp.351-360
• /
• 2010

In 2010, the number of registered vehicles reached almost at 17.48 millions in Korea. This dramatic increase of vehicles influenced to increase the number of traffic accidents which is one of the serious social problems and also to soar the personal and economic losses in Korea. Through this research, an enhanced intersection hazard prediction model by combining Genetic Algorithm and Artificial Neural Network will be developed in order to obtain the important data for developing the countermeasures of traffic accidents and eventually to reduce the traffic accidents in Korea. Firstly, this research has investigated the influencing factors of road geometric features on the traffic volume of each approaching for the intersections where traffic accidents and congestions frequently take place and, a linear regression model of traffic accidents and traffic conflicts were developed by examining the relationship between traffic accidents and traffic conflicts through the statistical significance tests. Secondly, this research also developed an intersection hazard prediction model by combining Genetic Algorithm and Artificial Neural Network through applying the intersection traffic volume, the road geometric features and the specific variables of traffic conflicts. Lastly, this research found out that the developed model is better than the existed forecasting models in terms of the reliability and accuracy by comparing the actual number of traffic accidents and the predicted number of accidents from the developed model. In conclusion, it is expect that the cost/effectiveness of any traffic safety improvement projects can be maximized if this developed intersection hazard prediction model by combining Genetic Algorithm and Artificial Neural Network use practically at field in the future.

• Journal of the Korea Institute of Building Construction
• /
• v.18 no.6
• /
• pp.611-619
• /
• 2018

The researches have only conducted regarding construction safety management and risk on interior construction site(workers) and is insufficient about the exterior construction site(third party). As a result, ordinary people who were near construction sites have injured and hold a negative view when they think about the construction industry because construction industry have been exposed to them having a high accidents rate through media. In addition, the importance of industrial disaster prevention is emphasized at this point in time, the overall safety management system should be constructed with considering construction site external(third human) for improving the negative image of the construction industry among ordinary people. Therefore, the objective of this study is to develop the quantification model of human accident utilizing the insurance claim payout occurred construction site exterior(third party). In the future, it can be used as a reference for developing the safety management checklist in construction site interior exterior and development for forecasting control system of human accident.

• Journal of Korean Society for Atmospheric Environment
• /
• v.19 no.5
• /
• pp.595-610
• /
• 2003

The statistical indexes such as RMSE (Root Mean Square Error), Mean Bias error, and IOA (Index of agreement) are used to evaluate 3 Dimensional wind and temperature fields predicted by operational meteorological model RAMS (Regional Atmospheric Meteorological System) implemented in CARIS (Chemical Accident Response Information System) for the dispersion forecast of hazardous chemicals in case of the chemical accidents in Korea. The operational atmospheric model, RAMS in CARIS are designed to use GDAPS, GTS, and AWS meteorological data obtained from KMA (Korean Meteorological Administration) for the generation of 3-dimensional initial meteorological fields. The predicted meteorological variables such as wind speed, wind direction, temperature, and precipitation amount, during 19 ∼ 23, August 2002, are extracted at the nearest grid point to the meteorological monitoring sites, and validated against the observations located over the Korean peninsula. The results show that Mean bias and Root Mean Square Error are 0.9 (m/s), 1.85 (m/s) for wind speed at 10 m above the ground, respectively, and 1.45 ($^{\circ}C$), 2.82 ($^{\circ}C$) for surface temperature. Of particular interest is the distribution of forecasting error predicted by RAMS with respect to the altitude; relatively smaller error is found in the near-surface atmosphere for wind and temperature fields, while it grows larger as the altitude increases. Overall, some of the overpredictions in comparisons with the observations are detected for wind and temperature fields, whereas relatively small errors are found in the near-surface atmosphere. This discrepancies are partly attributed to the oversimplified spacing of soil, soil contents and initial temperature fields, suggesting some improvement could probably be gained if the sub-grid scale nature of moisture and temperature fields was taken into account. However, IOA values for the wind field (0.62) as well as temperature field (0.78) is greater than the 'good' value criteria (> 0.5) implied by other studies. The good value of IOA along with relatively small wind field error in the near surface atmosphere implies that, on the basis of current meteorological data for initial fields, RAMS has good potentials to be used as a operational meteorological model in predicting the urban or local scale 3-dimensional wind fields for the dispersion forecast in association with hazardous chemical releases in Korea.