http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
염소 고함유시멘트의 페이스트 유동성과 모르타르 강도발현성에 미치는 무기질 혼화재의 영향
정찬일,박수경,이의학,이경희,Jeong, Chan-Il,Park, Soo-Kyung,Lee, Eui-Hak,Lee, Kyung-Hee 한국세라믹학회 2007 한국세라믹학회지 Vol.44 No.1
Fluidity, setting time, hydration heat, bond water ratio, compressive strength, SEM and BET of OPC were measured by adding 1.0 wt% KCl and replacing 20 wt% mineral admixture in order to examine effects of blast furnace slag (BFS), limestone powder (LSP), and fly ash (FA) on fluidity and strength development of the cement contained much chloride. In general, the cement contained much chloride was high in heat of hydration, short in its setting time, low in its fluidity and low in its strength at 28 days due to the rapid hydration in its initial stage. As a result of the experiment, it has been demonstrated that fluidity became improved but the compressive strength at 28 days was decreased as replaced LSP to the cement contained much chloride. the fluidity and compressive strength at 28 days was improved as replaced BFS, the initial compressive strength development was improved due to the activation of initial reaction by KCl. Fluidity, initial compressive strength and late compressive strength at 28 days of cement contained much chloride replaced 5 wt% LSP and 15 wt% BFS concurrently was better than OPC, but the hydration heat was lower.
정찬일,전진호,이진,김광태,이시영,천경두 대한이비인후과학회 2002 대한이비인후과학회지 두경부외과학 Vol.45 No.12
Background and Objectives:The traditional adenoidectomy using adenotome and adenoid curette could not guarantee a clear operative field due to bleeding. Also the traditional transoral adenoidectomy was not always efective in the complete removal of ess of using the electric suction coagulator for adenoidectomy. Materials and Method:This study was completed by 60 patients who underwent adenoidectomy or adenoidectomy with ventilation tube insertion. For adenoidectomy, 30 cases were performed using the variable sized suction coagulator via nasal cavity and the other 30 were performed using the adenotome and adenoid curette with oral approach. The amount of intraoperative bleeding and the duration of surgery were recorded. A preoperative and Results:Intra-operative bleeding amount was less and les time was needed in suction coagulator method (p<0.05). There was no statistical difference in the postoperative endoscopic grade, adenoid nasopharyngeal ratio, and the shortest nasopharyngeal diameter between two groups. The postoperative improvement of subjective symptoms was not diferent. Conclusion:Adenoidectomy using tion and complete hemostasis. Especialy, the authors think that this method is useful for the removal of superior part of adenoid and peritubal adenoid tissue. (Korean J Otolaryngol 2002 ;45 :1167-71)
설명 가능한 인공지능 기반의 프로세스 마이닝 분석 자동화 연구
정찬일,이후진 대한전자공학회 2019 전자공학회논문지 Vol.56 No.11
Process mining is one of the business management techniques for business innovation. While it automates the creation of execution process models in the information system log, it still relies on the analyst's relevant experience and knowledge of the business domain in analyzing the process content and the root cause of the problem. Since most of the derived process models are very complicated, an accurate analysis is impossible due to the limitations of human cognition ability, and most of the process mining algorithms go through an abstraction process. There is a possibility that distortions may occur in this process, or important issues may be missed. Hence, in this study, we attempted to predict the process results by using the process prediction model incorporating a LSTM algorithm of machine learning, and then to apply an explainable artificial intelligence(XAI) technique. Thus, this study has drastically reduced the reliance on the human intuition and knowledge in the root cause analysis of process mining, enabling the quantitative and rapid analysis. 업무 혁신을 위한 경영기법 중 하나인 프로세스 마이닝은 정보 시스템 로그에서 실행 프로세스 모델을 작성하는 것을 자동화하였으나, 프로세스 내용을 분석하여 문제에 대한 원인을 분석하는 영역에서는 여전히 분석가의 관련 경험과 업무 도메인에 대한 지식에 많이 의존하고 있다. 특히 대부분의 도출된 프로세스 모델이 매우 복잡하기 때문에 사람의 인식 능력의 한계로 인해 정확한 분석이 불가능하여 프로세스 마이닝 알고리즘에서는 추상화 과정을 거치는데, 이 과정에서 왜곡이 발생하거나 정작 중요한 이슈를 놓치게 될 가능성이 있다. 본 연구에서는 먼저 머신러닝 모델 중 LSTM 알고리즘을 이용한 프로세스 예측 모델을 개발하여 프로세스의 수행결과를 예측하였고 다시 이 예측모델에 설명 가능한 인공지능 기법을 적용하여 분석함으로써 최종적으로 프로세스 수행결과에 가장 큰 영향을 미치는 원인을 탐색하는 작업을 자동화하였다. 이를 통해 프로세스 마이닝에 대한 원인 분석 과정에서 인간의 직관과 업무지식에 대한 의존을 획기적으로 낮추었으며 정량적이고 신속한 프로세스 분석이 가능하게 되었다.