스파크 프레임워크에서 연속집계 질의처리를 위한 자원공유기법

정원일 한국지식정보기술학회 2020 한국지식정보기술학회 논문지 Vol.15 No.5

Recently, as the real-time data generated by sensors in connection with location information rapidly increases in the IoT environment, studies on various services using a big data platform have been conducted. In this big data platform, research on continuous query processing including data aggregation operation has been proposed to effectively analyze sensor big data constantly flowing for multi-user needs. However, the existing resource-sharing-based aggregation operation has a problem of concurrent execution of time-based and tuple-based queries and an increase in memory usage and deterioration in query processing performance due to duplicated maintenance of aggregate information. Therefore, in this paper, we propose a resource sharing method to effectively support continuous aggregation query processing based on Spark, an in-memory based distributed processing framework. The proposed method minimizes the increase in the cost of processing aggregate information through linear resource sharing based on summary information for the scope of query processing, and reduces the memory usage required for continuous aggregate query processing. Also, our method improves query processing performance by preventing duplicate generation of aggregate information. The proposed approach is implemented based on Spark framework to ensure real-time performance of continuous aggregate query processing for big data. Finally, through performance evaluation, it is shown that the proposed resource sharing technique can be effectively used for continuous aggregation query processing.

다중이용업소 화재에서 소방시설의 화재영향인자와 작동 효용성 분석

정원일 한국지식정보기술학회 2022 한국지식정보기술학회 논문지 Vol.17 No.3

Multiplex available premises accessible to everyone are prone to cause casualties and property damage due to fire. Due to these fires, casualties were high in entertainment pubs and accommodations for examiners, and large property damage was shown in general restaurants and game rooms. Previous researches were conducted on improvement of fire risk assessment indicators based on risk characteristics and safety management of multiplex available premises, development of devices for effective extinguishing the fire and evacuation, and analysis of fire factors based on big data. This study analyzed the effects of fire factors such as building information and damage information on firefighting facilities using big data-based analyzing technologies and the operation status of firefighting facilities at actual fire sites in multiplex available premises. Through this, it was intended to provide a scientific basis for establishing policies on fire prevention and response by grasping the utility of fire facilities in multiplex available premises. In the experimental results, the importance of firefighting facilities was shown in the order of the ground floor level, the burnt area, the basement level, the ​​soot area, and the property damage. In addition, firefighting facilities showed an operation rate of 64.1% in multiplex available premises fires, and in the operation status by classification, fire extinguishing facilities, alarm facilities, evacuation facilities, fire extinguishing water facilities, and fire extinguishing activity facilities were analyzed in order. From analyzed results, we suggested that it is necessary to strengthen the installation of firefighting facilities and to prepare institutional measures for the smooth operation of firefighting facilities at fire sites in order to minimize the social damage caused by fires.

경주 남산의 자연환경과 문화재 훼손 가능성에 관한 연구

정원일 대한건설환경학회 2003 건설환경연구 Vol.2 No.2

안전보건교육 의무화와 산업재해

정원일(Won Il Jeong)(鄭元一),이명선(Myung Sun Lee)(李明善),전용일(Yong Il Jeon)(田庸一) 위기관리 이론과 실천 2013 Crisisonomy Vol.9 No.10

본 연구는 사업장측면에서 외생적인 원인에 의한 근로자의 인적자본 손실인 산업재해가 발생하는 원인을 고찰하고, 산업재해를 감소시키기 위한 방안으로 산업안전보건교육이 산업재해 발생에 미치는 영향에 대하여 살펴본다. 이를 위하여 본 연구에서는 2009년 기준 산업안전보건동향조사 원자료를 사용하여, 기업의 산업안전보건교육 실태를 살펴보았다. 기업의 업종, 규모, 근로자 구성 특성, 산업안전보건과 관련한 정부지원정책, 산재 노출 가능 시간 등의 설명 변수를 활용하여 실증분석을 진행하고 산업안전보건교육의 효과성을 검증하였다. 재해자수에 영향을 미치는 요인을 포아송 회귀분석으로 분석한 결과, 서비스사업장에서의 비사무직 근로자를 대상으로 하는 산업안전보건 정기교육은 재해자수를 감소시키는 데 유의한 역할을 하고, 제조업 사업장에서의 작업내용 변경시에 실시되는 보수안전교육은 재해자수가 감소한다는 결론을 얻을 수 있다. 하지만, 총괄적인 측면에서 산업안전보건교육이 큰 효과를 거두지 못하고 있어서, 산업안전보건교육의 제도적, 운영적 측면에서 개선이 필수적인 것으로 보인다. This study investigates the effects on how the education of occupational safety and health education has influenced on industrial accidents. Using the 2009 Korean Occupational Safety and Health Survey data, we test the effectiveness of occupational safety and health education, classified by company size, workers' characterisitcs, government policies related with occupational safety and health, and other individual-level variables such as an exposure time. Our empirical analysis indicates that the occupational safety and health education on the non white-color job-workers in service sectors has the positive effect on the number of industrial accident worker and also the education for working condition changes can reduce the number of industrial accidents in the manufacturing sectors. Nevertheless, most of occupational safety and health education in Korean cannot reduce the industial accidents dramatically, and thus institutional and operational improvement is required for Korean occupational safety and health education systems.

아방가르드 담론에 관한 연구 : 포스트모던시대의 네오아방가르드를 중심으로

정원일 평택대학교 1997 論文集 Vol.9 No.1

Identification of Focus Information with Contextual and Lexical Cues by Korean Learners of English: Evidence from Reaction Times Analyzed by Linear Mixed-Effect Models

정원일,박범식 한국영어학회 2019 영어학 Vol.19 No.4

The current study examined the interaction between contextual and lexical focus cues in sentence processing of the identification of focus information in double object construction by Korean Learners of English. The study conducted three self-paced reaction time experiments and the results were analyzed by linear mixed-effects models (LMM). We found Korean speakers preferred the context specifying focus on direct object when processing interrogative contexts specifying focus on either the indirect or direct object in double object constructions. We also found that unlike English speakers, Korean speakers had difficulty in processing the elliptical remnant, cued by lexical focus. We suggest that the processing difficulty might be ascribed to Korean speakers’ insensitivity to the effect of relative positions of lexical focus cues.

거주지 화재에서 인명피해에 영향을 미치는 화재인자 영향평가에 관한 연구

정원일 한국지식정보기술학회 2021 한국지식정보기술학회 논문지 Vol.16 No.3

In recent years, fires in a diverse residential environment have resulted in continued social losses, including casualties. According to Korean National Fire Agency (NFA)’s Statistical Yearbook(2020), compared to all fires, the frequency of fires in residential areas such as houses and apartments is 26%, but the number of casualties is 46%. Accordingly, studies such as fire prediction, fire safety performance evaluation plan and fire vulnerability factor analysis to minimize damage from residential fires, and fire-fighting laws according to changes in fire environment were conducted. In addition, studies are being actively conducted to analyze fire risk factors based on big data in accordance with the growing fire trend and establish fire fighting policies based on this. We propose big data-based impact analysis and evaluation on fire factors that affect human damage due to fire, targeting fire data from public data of the NFA, and show policy measures for major fire factors. In this study, the data distribution characteristics of fire factors were analyzed, and those of fire were selected for the assessment of the impact on human damage from the results of the significance probability and correlation analysis between human damage and fire factors. In addition, the importance was calculated through a model for evaluating the impact of fire factors on human damage using the random forest algorithm. As a result of the experiment, participated equipment, participating personnel, property damage, etc. were analyzed as the main factors that have an influence in relation to human damage.

Neural Network Language Models as Psycholinguistic Subjects: Focusing on Reflexive Dependency

정원일,박명관 대한언어학회 2022 언어학 Vol.30 No.4

The purpose of this study is to investigate the reflexive-antecedent dependency resolution accompanying the wh-filler-gap dependency resolution in neural network language models (LMs)’ sentence processing, comparing the processing result of LMs to the one of humans. To do so, we adopt the psycholinguistic methodology that Fraizer et al. (2015) used for humans. The neural-network language models employed in this study are four LMs: the Long Short-Term Memory (LSTM) trained on large datasets, the Generative Pre-trained Transformer-2 (GPT-2) trained on large datasets, an LSTM trained on small datasets (L2 datasets), and the GPT-2 trained on small datasets (L2 datasets). We found that only the LMs trained on large datasets were sensitive to the dependency between a reflexive and its antecedent matching in gender, but all of the four neural LMs failed to learn reflexive-antecedent dependency accompanying wh-filler-gap dependency. Furthermore, we also found that the neural LMs have a learning bias in gender mismatch.

CER 선형결합을 통한 비용추정 모델 개발

정원일,이용복,김동규,강성진 대한산업공학회 2012 산업공학 Vol.25 No.3

The acquisition cost of defense weapon system has been continuously increasing because of art-of-technology of it. This phenomenon requires efficiency and transparency in the weapon system acquisition process through cost estimation. Therefore cost estimation is very important to the government acquisition programs to support decisions about funding and to evaluate resource requirement as a key decision point. The Commercial parametric cost estimating models have been using extensively to obtain appropriate cost estimates in early acquisition phase. These models have many restrictions to ensure the cost estimating result in Korean defense environment because they are developed based on foreign R&D data. Also estimation results are different from Korean defense industry accounting system. So, some studies have been tried to develop a CER (Cost Estimation Relationship) based on the Korean historical data. However, there are some restrictions to improve the predictability and ensure the stability of the developed singular CERs which consider the following data characteristics individually. The the abnormal conditions of data that is multicollinearity, outlier and heteroscedasticity under rack of the number of observations. In this paper, a CER’s Linear Combining Model is proposed to overcome those limitations which guarantee more accurate estimation (25.42% higher precision) than other singular CERs. At least, this study is meaningful as a first attempt to improve the predictability of CER with insufficient data. The methodology suggested in this study will be useful to develop a complex Korean version cost estimating model development in future.

전통 도시 숲 형성배경에 관한 연구

정원일,윤석 경주대학교 건설환경연구소 1999 建設環境論叢 Vol.- No.1

In the Shilla age, Kyongju was to be formed urban woods on the great area. But Urban development converted these traditional urban woods into concrete and farm. This study sought a problems and proposed the direction of improving in traditional urban woods of Kyongju. The results of this research were as follows: (ⅰ) the second pine trees were nothing, (ⅱ) foreign trees were greatly planted, (ⅲ) injudicious development was done in the inside and outside of ones, (ⅳ) it was destroyed by people's digging fruits and berries, and (ⅴ) it was to be absent of efficient management of an administrative organizatio. Traditional urban woods is an area of historical, cultural, and social activity. Therefore, it is needed to practice and plan for preserving the woods.