• Fire Science and Engineering
• /
• v.28 no.2
• /
• pp.82-90
• /
• 2014

According to recent government's 3.0 operating paradigm for the opening and sharing of public information, relationship between humidity, temperature and fire occurrence were analyzed using the data in National Weather Service and National Emergency Management Agency. In order to analyze the relationships between humidity, temperature and fire occurrence, hourly frequency of fire occurrence compared with humidity and temperature ranges was suggested as an analysis method. Tendencies of fire occurrence frequencies were examined through this and characteristics of detailed attributes could be statistically identified. Results about hourly frequencies of fire occurrence by classifying the humidity ranges in each region showed increasing frequencies in all areas where the humidity is lower. Hourly frequencies of fire occurrence according to temperature ranges were identified to be similar in each area as well. In addition, characteristics of objects' attributes were analyzed including types of fire, igniting source of fire, initial complex, reasons of fire occurrence, and distinctive directions were suggested. Suggested method in this paper could be practically used when suggesting the frequency in each category in fire occurrence statistics of National Fire Information System.

• The Journal of the Korea Contents Association
• /
• v.15 no.9
• /
• pp.351-363
• /
• 2015

Discourse analysis is useful for examining discourse characteristics of dementia patients and differentiating them from normal adults. However, inconsistent tasks, transcription methods and measures have been utilized, thus analyses were inevitably unsystematic. The purpose of this study was to investigate tasks, transcription methods and measures of discourse studies in dementia. Studies on dementia discourses written in English were collected via Scopus database containing abstracts and citations from 2005 to 2014, and were analyzed in detail. From the results, the most frequently used task was 'storytelling-topic'. It was observed that use of consistent transcription conventions rendered more reliable transcription data. Measures could be categorized into coherence, cohesion, productivity, propositional syntactic lexical analysis, and fluency, and each category belonged to discourse processing levels (i.e., processing of conceptual networks, propositions, and natural language units). In conclusion, future research warrants not only utilization of discourse tasks depending on dementia types and features, but also application of consistent transcription conventions. Additionally, measures based on the discourse processing theory would enable us to systematically interpret discourse studies.

• Proceedings of the Korea Institute of Fire Science and Engineering Conference
• /
• 2010.10a
• /
• pp.349-353
• /
• 2010

본 연구에서는 통계분석방법 가운데 하나인 의사결정분석방법을 활용하여 소방방재청 국가화재정보시스템의 2007년부터 2009년까지 강원지역에서 발생한 화재발생 데이터를 대상으로 화재발생에 대한 유형을 분석하였다. 이와 같은 분석결과는 체계적이고 효율적인 소방정책을 수립하기 위한 기초 자료로 활용될 수 있을 것으로 사료된다. 변수선정은 NFDS의 변수들 가운데 화재유형분석에 영향을 줄 것으로 판단되는 변수들만을 대상으로 요인변수들을 발화환경, 화재원인, 진화요인 등 3개의 그룹으로 분류하였으며, 목적변수로는 화재피해와 화재건수를 선정하였다. 또한 NFDS의 제공 데이터들 이외에도 분석의 신뢰도와 정확도를 높이고자 통계청에서 제공하는 2007년과 2008년 외부데이터를 포함시켰다. 분석방법은 대분류 차원의 화재유형별 분석을 수행하고자 인명피해, 재산피해, 화재건수 등 3개의 목적변수를 대상으로 변수들의 영향력을 고찰하였다.

• Fire Science and Engineering
• /
• v.32 no.6
• /
• pp.91-99
• /
• 2018

Recently, the spread of ESS in Korea has increased and a fire accident has also occurred. By July 2018, there were a total of 7 cases. All 7 cases were ESS systems consisting of lithium-ion batteries and were burned down. Both the automatic fire extinguisher and the fire department were not able to digest. In this paper, the characteristics of ESS fire are analyzed based on recent ESS fire situation and field investigation, and the cause of fire is divided into environmental, electrical and thermal factors. As a result, it was found that the ESS fire was correlated with the installation environment of the system. In the domestic and overseas lithium ion battery test standard and ESS facility standard survey, the trends and differences of domestic and overseas facilities standards were identified. Based on the fire status and field investigationy, and domestic and overseas facility standard survey, measures were suggested to prevent and prevent the spread of fire in ESS fire.

• Fire Science and Engineering
• /
• v.19 no.1 s.57
• /
• pp.59-69
• /
• 2005

There were little available data of vehicular incendiary fires. We investigated fires by inflammable gasoline and papers. cloths in the interior and exterior of vehicles. Fires in the interior and exterior of vehicles were carried out under the several conditions by incendiary methods and vehicle's windows. We knew that dominant factor of combustion is direction of wind at that time in vehicle's fire. And we knew that the other mark of combustion pattern in vehicle's fire could be a clue for solving the problem in discerning the place to which we sotted fire between the general fire and arson.

• Proceedings of the Korean Information Science Society Conference
• /
• 2004.10a
• /
• pp.811-813
• /
• 2004

휴대전화에서 전달되는 음성데이터들이 전달되는 과정에서 잡음 등의 외부 요인으로 인하여 데이터에 손실이 생기는 문제가 발생한다. 이렇게 전달된 음성데이터가 음성 인식기를 통과하면 바로 음성 인식기를 통과했을 때 보다 인식률이 낮아진다. 본 연구에서는 음성인식 알고리즘을 이용하여 홈 네트워크를 제어하는데 있어서 음성 인식율을 향상시키기 위해서 반복적으로 음성 데이터를 입력받아. 이를 유사율 알고리즘을 적용시켜 추출 된 여러 개의 데이터(text)를 이미 구축된 홈 네트워크 용어 관련 사전에 등록된 단어와의 유사성을 검토하여 추출된 결과로 홈 네트워크를 제어하는 방안을 제안한다. 이 결과, 기존의 방법에 비해서 10% 정도의 인식률의 향상을 확인할 수 있었다.

• Annual Conference on Human and Language Technology
• /
• 2023.10a
• /
• pp.107-112
• /
• 2023

라벨 데이터 수집의 어려움에 따라 라벨이 없는 데이터로 학습하는 준지도학습, 비지도학습에 대한 연구가 활발하게 진행되고 있다. 본 논문에서는 그의 일환으로 Novel Intent Category Discovery(NICD) 문제를 제안하고 NICD 연구의 베이스라인이 될 모델을 소개한다. NICD 문제는 라벨이 있는 데이터와 라벨이 없는 데이터의 클래스 셋이 겹치지 않는다는 점에서 기존 준지도학습의 문제들과 차이가 있다. 제안 모델은 RoBERTa를 기반으로 두 개의 분류기를 추가하여 구성되며 라벨이 있는 데이터셋과 라벨이 없는 데이터셋에서 각각 다른 분류기를 사용하여 라벨을 예측한다. 학습방법은 2단계로 먼저 라벨이 있는 데이터셋으로 요인표현을 학습한다. 두 번째 단계에서는 교차 엔트로피, 이항교차 엔트로피, 평균제곱오차, 지도 대조 손실함수를 NICD 문제에 맞게 변형하여 학습에 사용한다. 논문에서 제안된 모델은 라벨이 없는 데이터셋에 대해 이미지 최고성능 모델보다 24.74 더 높은 정확도를 기록했다.

• Smart Media Journal
• /
• v.13 no.6
• /
• pp.54-61
• /
• 2024

Recently, due to the increase in domestic and international online e-commerce platforms and the increase in container traffic at domestic ports, the operating ratio of large trucks has increased, and the number of truck fires is continuously increasing. In particular, spontaneous combustion is the most common cause of truck fires. Various academic approaches have been attempted to prevent truck fires, but due to the lack of research on the spontaneous tire ignition phenomenon that occurs during braking, this research directly designed and manufactured an experimental device to establish an environment similar to the braking system of a truck. A non-contact temperature sensor was installed on the brake device of the experimental device to collect temperature data generated from the brake device. Based on the data collected from the temperature sensor of the brake device and the temperature sensor on the tire surface, the ARIMA model among the time series prediction models was used to Appropriate parameters were selected to suit the temperature change trend, and as a result of comparing and analyzing the measured and predicted data, an accuracy of over 90% was obtained. Based on this, a plan was proposed to reduce the rate of fires in trucks by providing real-time warnings and support for truck drivers to respond to overheating phenomena occurring in the braking system.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.39 no.6
• /
• pp.351-361
• /
• 2021

The seriousness of fire is rising because fire causes enormous damage to property and human life. Therefore, this study aims to predict various risk factors affecting fire by fire type. The predictive analysis of fire factors was carried out targeting Gyeonggi-do, which has the highest number of fires in the country. For the analysis, using machine learning methods SVM (Support Vector Machine), RF (Random Forest), GBRT (Gradient Boosted Regression Tree) the accuracy of each model was presented with a high fit model through MAE (Mean Absolute Error) and RMSE (Root Mean Squared Error), and based on this, predictive analysis of fire factors in Gyeonggi-do was conducted. In addition, using machine learning methods such as SVM (Support Vector Machine), RF (Random Forest), and GBRT (Gradient Boosted Regression Tree), the accuracy of each model was presented with a high-fit model through MAE and RMSE. Predictive analysis of occurrence factors was achieved. Based on this, as a result of comparative analysis of three machine learning methods, the RF method showed a MAE = 1.765 and RMSE = 1.876, as well as the MAE and RMSE verification and test data were very similar with a difference between MAE = 0.046 and RMSE = 0.04 showing the best predictive results. The results of this study are expected to be used as useful data for fire safety management allowing decision makers to identify the sequence of dangers related to the factors affecting the occurrence of fire.

• Journal of The Korean Association For Science Education
• /
• v.32 no.4
• /
• pp.555-569
• /
• 2012

Many science education research and practices are recently emphasizing the importance of collaborative learning. This study also understands learning in aspects of socio-cultural context, and regarded the creation of meaning in a same-age group as an important learning process. This is most especially true in the premise that the formation of roles in a collaborative learning is important for successful interactive learning. This study aims to find out how roles form in a group. For this purpose, university students participating in a group discussion activity about energy flow and circulation of material were selected as research participants. Discussions among the nine students in one group consisted of cognitive conversations on the topic and operational conversations for preparing a presentation. Video-clips of the discussions were made and transcribed. For the analysis, we developed a framework that includes four interaction functions (cognitive, organizational, meta-cognitive, operational), four action elements (question, simple answer, providing opinion, response to opinion), and two to four intention elements by each action elements. As a result, a total of nine roles were revealed through the interaction function and element; cognitive questioner, operational questioner, simple answerer, operational suggester, organizational commander, operational commander, cognitive explainer, terminator, reflective thinker. These roles are re-classified into seven utterance patterns by the utterance order and object, and they were categorized into three role groups (facilitating interaction, sustaining interaction, finishing interaction). The result means that role formation and function can have influence on learning and interaction. This study is meaningful to the suggestion to collaborative learning including project-based learning, investigation, club activity, and for the re-illumination of the role in an aspect of the interaction.