인터넷과 스마트폰의 보급과 함께 사람들은 관심 있는 정보를 포털사이트에서 실시간으로 검색할 수 있게 되었다. 그에 따라 빅데이터가 비정형적으로 형성되고 이를 이용하기 위한 연구 분야가 발전하고 있다. 최근 식품분야 에서도 빅데이터를 활용한 연구의 필요성이 꾸준히 제기 되고 있다. 본 연구는 국내 최대 규모의 포털 사이트인 네이버의 검색 통계를 활용하여, 사람들의 연관 검색어 빈도에 따른 식품안전 사고 원인 식품과 식중독 원인 병원 체들의 관계를 분석하였다. 특히 식중독과 관련된 식품과 식중독 원인 병원체의 포털사이트 키워드 검색 건수와 해당 원인 식품 및 병원체의 뉴스검색 건수를 분석하였으며, 최종적으로 키워드 검색 건수와 실제 식중독 발생 시기의 상관관계를 분석하였다. 분석결과, 일반적으로 식중독 발생 직후 관련 키워드 검색 건수가 증가하는 것으로 나타 났으며, 특별하게 굴과 같은 계절성 식품의 경우 키워드 검색 건수 증가 이후에 식중독 사고가 발생한 것으로 나타났다. 따라서 본 연구결과를 통해 검색 건수가 증가하는 식중독 세균이나 식품들을 식중독 사전 예방적 차원의 정보로 활용할 수 있다면 식중독 사고 발생 가능성을 낮출 수 있을 것으로 사료된다.

This paper reviewed the relationship among job characteristics, psychological stress response, and accidents, and the mediating effect of psychological stress response between job characteristics and accidents. Based on the responses from 176 employees, the results of multiple regression analysis appeared as follow; 1) autonomy, importance and feedback effect negatively on psychological stress response. 2) psychological stress response effects positively on accident frequency. 3) psychological stress response mediates fully between autonomy and accident. 4) psychological stress response mediates partly between feedback and accident.

본 연구에서는 고속도로를 대상으로 각각의 구간에 대한 선형유형을 구분하여 사고빈도모형을 개발하였다. 현재 사고빈도모형 부문의 연구는 주로 고속도로 구간 전체를 대상으로 한 연구가 대부분이기 때문에 거시적인 측면에서 사고빈도모형이 개발되었다고 할 수 있으며, 이에 따라 각각의 구간특성이 정확히 반영되지 않은 상태에서의 사고빈도를 예측하였다고 볼 수 있다. 본 연구에서는 이러한 문제점을 인식하여, 동질구간 분할법을 사용하여 고속도로의 평면선형을 직선부, 곡선부, 연속곡선부로 구분하였고, 이를 군집분석을 통하여 직선부와 곡선부의 유형을 구분하여 고속도로의 각각의 구간별 특성을 반영한 사고빈도모형을 구축하였다. 본 연구 결과는 고속도로 각 구간의 사고빈도를 예측하는데 있어 더 정확하고 합리적인 결과를 도출해 낼 것이라 판단한다.

In this study, with the objects of 173 institutions of university and research institutions, investigation was carried out through visit investigation ar d questionnaire investigation. Frequency and severity of occurrences was acquired mainly through universities and research institutions which have reported occurrences of accidents to the Ministry of Education, Science, and Technology. Quantitative danger evaluation was conducted through frequency and strength of occurrences. Case study was carried out by selecting bio science laboratory of university and as the result of this study, degree of danger among the inspection objects of bio science laboratories, laboratory work place was evaluated as the most dangerous with the scores of 155.00 and as the result of average danger level, compression gas cylinder was evaluated as the most dangerous with the scores of 6.4000.

In this study, with the objects of 173 institutions of university and research institutions, investigation was carried out through visit investigation and questionnaire investigation. Frequency and strength of occurrences was acquired mainly through universities and research institutions which have reported occurrences of accidents to the Ministry of Education, Science, and Technology. Quantitative danger evaluation was conducted through frequency and strength of occurrences. Case study was carried out by selecting bio science laboratory of university and as the result of this study, degree of danger among the inspection objects of bio science laboratories, laboratory work place was evaluated as the most dangerous with the scores of 155.00 and as the result of average danger level, compression gas cylinder was evaluated as the most dangerous with the scores of 6.4000.

  We use influence diagrams to describe event trees used in safety analyses of low-probability high-risk incidents. This paper shows how the branch parameters used in the event tree models can be updated by a bayesian method based on the observed counts o

Several researchers have discussed which element is preferably chosen as a vehicle in metonymic constructions Langacker (1993) claims that, all other things being equal, these cognitive principles govern the choice of metonymic vehicle: human over non-human, whole over part, concrete over abstract, visible over non-visible. This claim has been made, but no previous case studies regarding these principles have been conducted. To remedy this gap, this paper examines 510 instances of PART FOR WHOLE metonymy in the names of events and accidents in Korean newspapers from 2010 through 2019. The result shows that AGENT is most frequently utilized to refer to whole events, supporting one principle of Langacker, human over nonhuman. CAUSE, PATIENT, and PLACE follows AGENT. INSTRUMENT is not very frequently used unless it achieves its salience through contexts. TIME is observed only once. Considering that TIME was used frequently to name past events, as in the 4.19 Revolution, 5.18 Democratic Movement, the 4.3 Events, the infrequent use of TIME reflects a cultural change in its cognitive salience.