초록문의 abstract@kcsnet.or.kr

결제문의 member@kcsnet.or.kr

현재 가능한 작업은 아래와 같습니다.
  • 09월 01일 18시 이후 : 초록수정 불가능, 일정확인 및 검색만 가능

제118회 대한화학회 학술발표회, 총회 및 기기전시회 안내 QSPR modeling for screening high energy molecules(HEMs) using machine learning methods

2016년 8월 31일 09시 57분 22초
PHYS1-5 이곳을 클릭하시면 발표코드에 대한 설명을 보실 수 있습니다.
목 15시 : 05분
물리화학 - Machine Learning and Chemistry
저자 및
이성광*, 박한웅, 김현정, 이민지, 이병훈, 서현일1, 조수경2, 김한조3, 윤정혁4
한남대학교 화학과, Korea
1(주)퀀텀소프트, Korea
2국방과학연구소 4본부-2부, Korea
3CJ헬스케어 신약연구센터, Korea
4(주)파로스아이비티 -, Korea
High energetic molecules (HEMs) are widely used for various civil, military and architecture applications. Generally, the promising HEMs should have the characteristics of enhanced detonation performance and low impact sensitivity. Removal of poor candidate HEMs before investing in synthesis and testing will go far toward developing the promising HEMs. Therefore, the importance of predictive modeling for searching potential HEMs has also increased as it minimizes extremely cost/time-consuming and risky step of synthesis as well as experiments of untested HEMs. The main parameters to evaluate the sensitivity and performance of HEMs are impact sensitivity, which expresses safe quality of HEMs to external stimuli, and detonation velocity or pressure, which were calculated from heat of formation and density by Kamlet-Jacobs equation. In this study, we will explain predictive QSPR modeling for screening insensitive HEMs, including representation of HEM structure, and supervised machine learning. The QSPR models were evaluated according to the OECD principles for QSAR/QSPR validation, and compliance with all five principles was established. Finally, we present newly developed softwares; that is, MS-HEMs 2 and PharosQSAR. MS-HEMs2 has been developed in response to need for information management and rapid prediction of HEMs properties. PharosQSAR is designed to build, utilize and import/export QSPR models. These softwares enables us to facilitate rational design of HEMs and to share QSAR/QSPR models was published in journal.