• The Korean Journal of Applied Statistics
• /
• v.32 no.6
• /
• pp.851-865
• /
• 2019

Discovering various features from one time series is complicated. In this paper, we introduce a multi aggregation prediction algorithm (MAPA) that uses the concepts of temporal aggregation and combining forecasts to find multiple patterns from one time series and increase forecasting accuracy. Temporal aggregation produces multiple time series and each series has separate properties. We use exponential smoothing methods in the next step to extract various features of time series components in order to forecast time series components for each series. In the final step, we blend predictions of the same kind of components and forecast the target series by the summation of blended predictions. As an empirical example, we forecast traffic accident counts using MAPA and observe that MAPA performance is superior to conventional methods.

• International Journal of Highway Engineering
• /
• v.11 no.3
• /
• pp.85-95
• /
• 2009

According to the accident statistics published by the National Police Agency in 2007, the number of commercial vehicle accidents explains 3.5 percent of the total number of traffic accidents of the year. Compared to other types of vehicles commercial vehicles may provide more serious damages to both driver himself and passengers. Thus, they generate more serious social and economic problems. There have been various forms of systems such as a digital speedometer or a black box to meet the social requirement for reducing traffic accidents and improving safe driving. However, since the current systems are based on the data often accidents happened, there are lots of limitations to control drivers in real-time. Also, the current speedometers provide drivers with only speeds of vehicles and RPM information regardless of actual dangerous drive behaviors. Therefor, they lack of the effectiveness in terms of safety. In this research, real-time information systems for improving driver safety based on automatic risky driving behaviors, and thresholds to determine risky driving patterns were studied.

• International Journal of Highway Engineering
• /
• v.10 no.4
• /
• pp.117-125
• /
• 2008

Recently, users of bicycles as a green mode are rising sharply, but there are few efforts to increase bicycle safety and to decrease the number of accidents between vehicles and bicycle users. Therefore, this research seeks to improve bicycle safety at intersections by analysing causation factors of bicycle accidents pattern and providing optimal intersection design guidelines. For this study, real bicycle accident data of fifty six occurred in the Incheon metropolitan area in the year of 2005 were collected and field surveys were conducted. In addition, this research developed a bicycle accident pattern model with using multinomial logit model. The model results showed that presence of fatality, presence of traffic islands on the minor road, highway_type, weather, presence of traffic bus_stop on the major road, minor road lane width, victim induce factor are significant for bicycle safety.

• Journal of Korean Society for Atmospheric Environment
• /
• v.16 no.1
• /
• pp.23-35
• /
• 2000

PM10, which is below 10 ${\mu}{\textrm}{m}$ in a diameter, has a high deposition in the lung or the bronchus by breathing and is generally composed of a lot of organic matters, viruses, algae, mold, and metallic elements that are very toxic to people. This study identified the characteristics of concentration of PM10 and air-borne jmetallic elements produced in the industrial city, Ulsan, and analyzed the correlatuion between sources and generation patterns of PM10 and metallic elements. We classified the five areas(green, residential, heavy traffic, mechanic, and petrochemcal and non-ferrous metal) which might have different characteristics of sources of PM10 and metallic elements. The average concentrations of PM10 in the five areas were as follows(petrochemical and non-ferrous metal(99.9$\mu\textrm{g}$/㎥)>mechanic(77.5 $\mu\textrm{g}$/㎥)>heavy traffic(47.1 $\mu\textrm{g}$/㎥)>residential(39.3 $\mu\textrm{g}$/㎥)>green(32.8 $\mu\textrm{g}$/㎥)). Those of petrochemical and non-ferrous metal areas were higher than other areas. In this study, the average concentration trend of metallic elements contained in PM10 are shown as follows: Fe>Zn>Pb>Cu>Mn>Cr>As>Cd>Sn>Hg, respectively. The metallic elements identified in PM10 showed the highest concentration in the petrochemical and non-ferrous areas. Metal combinations showed that a high correlation among concentrations of heavy metals were as follows: As, Cd and Fe in the residential area; Zn, Mn, Cu and Pb in the mechanical area; and Zn, Cu, As, Pb in the petrochemical and non-ferrous industrial area.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.8 no.5
• /
• pp.48-56
• /
• 2009

Regulation of Visible Light Transmission Percentage (VLT%) has been argued because it was known that the degree of darkness of tinted vehicle can affect to driving maneuver. Previously, it was proven that low level of VLT affects capacity reduction. But, due to lack of field data they could not analyze the effect of Car-Following model parameters. This study focuses on the effect of a tinted vehicle on following traffic flow. RTK GPS receiving data through field experiment analyzed based on headway distance, acceleration noise, sensitivity, and reaction time. As a result of analysis through GM 1st Model and 3rd Model, influence of following vehicle vary inversely with VLT and risk according as tinting of lead vehicle is third vehicle bigger than second vehicle. Also the results patterns of GM 3rd model include distance-headway are same with GM 1st Model. In the further need to research for influence analysis of traffic flow stability by the level of VLT.

• Journal of Preventive Medicine and Public Health
• /
• v.37 no.3
• /
• pp.274-281
• /
• 2004

Objectives : This study was performed to investigate the injury rates and risk factors for preschool children in Daegu city and Kyungpook province. Method : A questionnaire survey about medically attended injuries during the preschool period was performed in nine primary schools located in Daegu city, Pohang city and Goryung County. The overall injury rate was estimated using person-year. The causes and patterns of the injuries, and their risk factors were examined. Result : A total of 469 medically attended injuries were reported in 330 of the 959 study subjects during the preschool period. The overall annual injury rate was 7.5 per 100 children. The injury rate increased sharply during the period from infant (2.4) to 1 year of age (7.5), and the peak injury rate (9.2) was reported for 5 year olds. The most common causes of injuries were falling (36.0%), followed by being struck by an object (23.7%), and traffic accidents (14.1%). Among the traffic accidents, 72.8% occurred while playing on the road, riding a bicycle or roller-skating. A proportional hazard model showed that males (hazard ratio=1.49, p<0.001 compared with female) and the mother's higher education level (hazard ratio of college or higher= 1.51, p=0.013; high school=1.32, p=0.085 compared with those of middle school or lower) were significant risk factors of childhood injury. Conclusion : The results of this study suggested that efforts for children's safety should be made, especially from the toddler stage, and in male children. To develop a more specific childhood injury prevention program, a surveillance system for injuries should be established. Further study of the relationship between mother's occupation and injury rates is also needed.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.33 no.1A
• /
• pp.59-66
• /
• 2008

This paper presents a spectrum sharing and allocation problem in multiple network operators(NOs). Most of previous works as to dynamic spectrum allocation(DSA) schemes have presented the DSA gain achieved by utilizing the time or regional variations in traffic demand between multi-network operators(NOs). In this paper, we introduce the functionalities required for the entities related with the spectrum sharing and allocation and propose a spectrum allocation algorithm while considering the long-term priority(or fairness) between NOs, the prioritybetween multiple class services, and the urgent bandwidth request. It was shown that the proposed algorithm with priorities between multiple class services could sustain better satisfaction ratio(SR) than the fixed allocation schemes or one without priority under different traffic patterns. Therefore, we conclude that the proposed DSA with priorities between multiple class services and network operators could provide the fairness between service classes of network operators and effective negotiation procedure for urgent bandwidth request by users.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.36 no.8B
• /
• pp.904-909
• /
• 2011

Cell zooming adjusts the cell range of base stations depending on traffic condition. The cell zooming can be implemented by the adjustment of antenna angles, the clustering of the base stations, and the cell relay. The base stations can adjust the cell range in term of energy efficiency, which can then reduce the overall energy consumption of cellular networks. There is, however, a trade-off between the energy savings and the blocking probability of user calls. A periodic scheme that manages the cell zooming of the base stations was proposed but it was inadequate for dealing with the dynamic nature of traffic patterns. This paper proposes a semi-periodic cell zooming scheme along with the algorithms that select such base stations and define the operation procedure. Simulation results show that the proposed method outperforms the existing scheme in terms of the energy savings without the degradation of the blocking probability.

• KIPS Transactions on Computer and Communication Systems
• /
• v.1 no.2
• /
• pp.109-118
• /
• 2012

As a result of the prevalence of smartphone, various mobile services became faster by LTE networks. Because many mobile devices are used more wireless services, heavy signaling traffic for authentication could be generated. Authentication is an important factor in wireless networks to identify devices; it is the start of wireless networks. This paper analyzes previous patterns for more effective authentication in accessing of another external networks. We propose a fast authentication scheme for minimizing of signaling cost between the authentication server and external networks. And we calculate the rate of authentication occurrence in LTE networks using mathematical modeling as well as the change of signaling cost for authentication in various network environments. Finally, we calculate the optimized number of authentication data and show the effectiveness for authentication signaling costs.

• International Journal of Highway Engineering
• /
• v.15 no.1
• /
• pp.127-134
• /
• 2013

PURPOSES: Speeding is often the primary contributor to fatal crashes. Surprisingly, driving behaviors are indirectly affected by personal factors such as law-abidance, risk sensitivity, and situational adaptability. This research aims to verify the effectiveness of pavement markings at reducing the speeds of high-risk drivers. The purpose of this study is to establish how drivers (including law-abiding or law-breaking, high-risk or low-risk) react to different pavement markings in a driving simulator. METHODS: The five different pavement markings were selected from markings used in other nations. The forty-two drivers were then surveyed, via questionnaires, and placed into the corresponding groups. Finally, statistical analysis was conducted to determine the extent of speed reduction for each pavement marking. RESULTS: Higher speeds were linked to the high-risk drivers. Furthermore, after analysis of the mean difference of average speeds by pavement marking, it was determined that Dragon's Teeth had the greatest speed reducing effect on these drivers. CONCLUSIONS: Perceptual countermeasures are unlikely to strongly affect high-risk drivers' perception of speed on the curves. This statistically demonstrates that Dragon's Teeth have a subtle effect on reducing speeds in the driving simulator. This study's significance lies in the improved understanding of high-risk drivers in terms of road facilities. It approaches the effects of various patterns of pavement markings for high-risk drivers.