• Proceedings of the KSR Conference
• /
• 2008.06a
• /
• pp.1524-1530
• /
• 2008

Generally a road vehicle's wrong entry into level crossing gives rise to hazardous events, the eventual collision with a approaching train depends on the effective operation of safety barriers such a abnormal condition detecting or emergency braking. In this paper, the risk assessment models developed for the level-crossing accidents will be introduced. The definition of hazardous events and the related hazardous factors are identified by the review of the accident history and engineering interpretation of the accident behavior. A probability of the hazardous events will be evaluated by the FTA, which is based on the accident scenario. For the severity estimation, the critical factors which can effect on the consequence will be reviewed during the ETA. Finally, the number of casualty for the public(vehicle drivers) and the train passengers are converted into an equivalent fatality.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.1
• /
• pp.169-189
• /
• 2015

Detection of anomalous events from video streams is a challenging problem in many video surveillance applications. One such application that has received significant attention from the computer vision community is traffic video surveillance. In this paper, a Lossy Count based Sequential Temporal Pattern mining approach (LC-STP) is proposed for detecting spatio-temporal abnormal events (such as a traffic violation at junction) from sequences of video streams. The proposed approach relies mainly on spatial abstractions of each object, mining frequent temporal patterns in a sequence of video frames to form a regular temporal pattern. In order to detect each object in every frame, the input video is first pre-processed by applying Gaussian Mixture Models. After the detection of foreground objects, the tracking is carried out using block motion estimation by the three-step search method. The primitive events of the object are represented by assigning spatial and temporal symbols corresponding to their location and time information. These primitive events are analyzed to form a temporal pattern in a sequence of video frames, representing temporal relation between various object's primitive events. This is repeated for each window of sequences, and the support for temporal sequence is obtained based on LC-STP to discover regular patterns of normal events. Events deviating from these patterns are identified as anomalies. Unlike the traditional frequent item set mining methods, the proposed method generates maximal frequent patterns without candidate generation. Furthermore, experimental results show that the proposed method performs well and can detect video anomalies in real traffic video data.

• Journal of Nuclear Fuel Cycle and Waste Technology(JNFCWT)
• /
• v.13 no.1
• /
• pp.1-10
• /
• 2015

In the safety assessment of the geological disposal system of the radioactive wastes, the abnormal scenarios, in which the system is impacted by the abnormal events, need to be considered in addition to the reference scenario. In this study, characterization and prediction of well intrusion as one of the abnormal events which will impact the disposal system were conducted probabilistically and statistically for the safety assessment. The domestic well development data were analyzed, and the prediction methodologies of the well intrusion were suggested with a computation example. From the results, the annual well development rate per unit area in Korea was about 0.8 well/yr/km² in the conservative point of view. Considering the area of the overall disposal system which is about 1.5 km², the annual well development rate within the disposal system could be 1.2 well/yr. That is, it could be expected that more than one well would be installed within the disposal system every year after the institutional management period. From the statistical analysis, the probabilistic distribution of the well depth followed the log-normal distribution with 3.0363 m of mean value and 1.1467 m of standard deviation. This study will be followed by the study about the impacts of the well intrusion on the geological disposal system, and the both studies will contribute to the increased reliability of safety assessment.

• The Sea:JOURNAL OF THE KOREAN SOCIETY OF OCEANOGRAPHY
• /
• v.26 no.4
• /
• pp.307-326
• /
• 2021

Real-time sea level observations from tide gauges include missing and erroneous values. Classification as abnormal values can be done for the latter by the quality control procedure. Although the 3𝜎 (three standard deviations) rule has been applied in general to eliminate them, it is difficult to apply it to the sea-level data where extreme values can exist due to weather events, etc., or where erroneous values can exist even within the 3𝜎 range. An artificial intelligence model set designed in this study consists of non-annotated recurrent neural networks and ensemble techniques that do not require pre-labeling of the abnormal values. The developed model can identify an erroneous value less than 20 minutes of tide gauge recording an abnormal sea level. The validated model well separates normal and abnormal values during normal times and weather events. It was also confirmed that abnormal values can be detected even in the period of years when the sea level data have not been used for training. The artificial neural network algorithm utilized in this study is not limited to the coastal sea level, and hence it can be extended to the detection model of erroneous values in various oceanic and atmospheric data.

• Asia Marketing Journal
• /
• v.16 no.1
• /
• pp.21-33
• /
• 2014

When the negative event is published, the company tends to go through the negative impact on the firm performance. Especially, with the SNS, the negative event is instantly spread on indefinite region so the impact seems bigger than the period before the SNS media appearance. It seems that everyone considers the SNS media impact on the firm performance quite big. However, there has been no empirical study on the impact comparison on the firm performance between pre and post SNS media occurrence periods. This study tries to empirically compare the impact of the negative event on the firm performance between pre and post SNS media appearance. Our study starts fromthe basic but not verified question; Does really the negative event have more negative impact in the post-SNS-occurrence period than in the pre-SNS-occurrence period? In order to examine the impact of the negative publicity on firm performance in two eras, pre and post SNS media appearance, we used CAR (Cumulative Abnormal Resturns) model. By using this model, we could verify the statistical significance of cumulative abnormal returns in market between before and after the events. For event samples, we focused on food manufacturers and collected the negative events from 1991 to 2003 for pre-SNS occurrence period, and from 2010 to 2013 for post-SNS occurrence period. Based on the listed food companies at KOSPI, we researched Naver News Library (newslibrary.naver.com) and Naver News (news.naver.com) for all the individual negative events published for both periods. Firm returns data were collected from TS 2000 (KOCO Info) and market portfolio data were collected from KRX Exchange. Through our empirical analysis, our finding is interesting to note that the type of events differently influences on the firm performance. With the SNS, the health-related events have influence on the firm performance 'after the event day' whereas the company behavior trust events have influence 'before the event day'. Our findings have implications for management. When a negative event directly related to or threatening customers or their life such as health, it is crucial to fix up the situation right after the event occurs. On the other hand, when a negative event is not publicly available information such as company behavior trust, it is important for marketers to strengthen the firms' trust reputation and control the bad WOM before the event.

• Journal of Korean Neurosurgical Society
• /
• v.43 no.6
• /
• pp.275-280
• /
• 2008

Objective : In order to assess the incidence of thromboembolic events and their clinical presentations, the present study prospectively examined routine brain magnetic resonance images (MRI) taken within 48 hours after a coil embolization of cerebral aneurysms. Methods : From January 2006 to January 2008, 163 cases of coil embolization of cerebral aneurysm were performed along with routine brain MRI, including diffusion-weighted magnetic resonance (DW-MR) imaging, within 48 hours after the embolization of the aneurysm to detect the silent thromboembolic events regardless of any neurological changes. If any neurological changes were observed, an immediate brain MRI follow-up was performed. High-signal-intensity lesions in the DW-MR images were considered as acute thromboembolic events and the number and locations of the lesions were also recorded. Results : Among the 163 coil embolization cases, 98(60.1%) showed high-signal intensities in the DW-MR imaging follow-up, 66 cases (67.0%) involved the eloquent area and only 6cases (6.0%) showed focal neurological symptoms correlated to the DW-MR findings. The incidence of DW-MR lesions was higher in older patients (${\geq}60$ yrs) when compared to younger patients (<60 yrs) (p=0.002, odd's ratio=1.043). The older patients also showed a higher incidence of abnormal DW-MR signals in aneurysm-unrelated lesions (p=0.0003, odd's ratio=5.078). Conclusion : The incidence of symptomatic thromboembolic attacks after coil embolization of the cerebral aneurysm was found to be lower than that reported in previous studies. While DW-MR imaging revealed a higher number of thromboembolic events, most of these were clinically silent and transient and showed favorable clinical outcomes. However, the incidence of DW-MR abnormalities was higher in older patients, along with unpredictable thromboembolic events on DW-MR images. Thus, in order to provide adequate and timely treatment and to minimize neurological sequelae, a routine DW-MR follow-up after coil embolization of cerebral aneurysms might be helpful, especially in older patients.

• The Journal of the Korea Contents Association
• /
• v.10 no.2
• /
• pp.99-110
• /
• 2010

An event means a flow which has a time attribute such as a symptom of patient. Stream data collected by sensors can be summarized as an interval event which has a time interval between the start-time point and the end-time point in multiple stream data environment. Most of temporal mining techniques have considered only the frequent events. However, these approaches may ignore the infrequent event even if it is important. In this paper, we propose a new temporal data mining that can find association rules for the significant temporal relation based on interval events in multidimensional stream data environment. Our method considers the weight of events and stream data on the sensing time point of abnormal events. And we can discover association rules on the significant temporal relation regardless of the occurrence frequency of events. The experimental analysis has shown that our method provide more useful knowledge than other conventional methods.

• Journal of Computing Science and Engineering
• /
• v.7 no.4
• /
• pp.272-284
• /
• 2013

In this paper, we propose a new framework for anomaly detection in medical wireless sensor networks, which are used for remote monitoring of patient vital signs. The proposed framework performs sequential data analysis on a mini gateway used as a base station to detect abnormal changes and to cope with unreliable measurements in collected data without prior knowledge of anomalous events or normal data patterns. The proposed approach is based on the Mahalanobis distance for spatial analysis, and a kernel density estimator for the identification of abnormal temporal patterns. Our main objective is to distinguish between faulty measurements and clinical emergencies in order to reduce false alarms triggered by faulty measurements or ill-behaved sensors. Our experimental results on both real and synthetic medical datasets show that the proposed approach can achieve good detection accuracy with a low false alarm rate (less than 5.5%).

• Journal of the Korea Society of Computer and Information
• /
• v.21 no.11
• /
• pp.151-158
• /
• 2016

This study analyzes the stock market reaction of disclosure of technological information using events which are collected in the Korean stock market for the thirteen-year period between January 2002 and December 2014. We find that abnormal return on the disclosure day of full sample firms is positive and statistically significant. However, abnormal return of high R&D intensity subsample is a larger positive number than that of the low one. Using a longer window, it shows that low R&D intensity negatively decreases the long term performance after the adoption of new technological information. The empirical evidence of the studying is expected to serve as a good judging guide-line for the investors.

• Journal of the Korea Society of Computer and Information
• /
• v.24 no.5
• /
• pp.59-66
• /
• 2019

According to the KT telecommunication statistics, people stayed inside their houses on an average of 11.9 hours a day. As well as, according to NSC statistics in the united states, people regardless of age are injured for a variety of reasons in their houses. For purposes of this research, we have investigated an abnormal event detection algorithm to classify infrequently occurring behaviors as accidents, health emergencies, etc. in their daily lives. We propose a fusion method that combines three classification algorithms with vision pattern, audio pattern, and activity pattern to detect unusual user events. The vision pattern algorithm identifies people and objects based on video data collected through home CCTV. The audio and activity pattern algorithms classify user audio and activity behaviors using the data collected from built-in sensors on their smartphones in their houses. We evaluated the proposed individual pattern algorithm and fusion method based on multiple scenarios.