View : 540 Download: 0

Full metadata record

DC Field Value Language
dc.contributor.advisor임용빈-
dc.contributor.author정종희-
dc.creator정종희-
dc.date.accessioned2016-08-25T04:08:28Z-
dc.date.available2016-08-25T04:08:28Z-
dc.date.issued2000-
dc.identifier.otherOAK-000000052322-
dc.identifier.urihttps://dspace.ewha.ac.kr/handle/2015.oak/177689-
dc.identifier.urihttp://dcollection.ewha.ac.kr/jsp/common/DcLoOrgPer.jsp?sItemId=000000052322-
dc.description.abstract축차적 설계가 검증할 화합물들이 증가하는 거대한 화학적 데이터 베이스에서 효율적이라는 것이 밝혀져 있다. 또한, 다중나무예측치들이 한 개의 나무 예측치보다 더 정확하다는것도 알려져 있다. 우리는 이러한 축차적 설계와 다중나무들을 이용하여, 다중나무를 갖고 축차적으로 접근하는 방법을 제안한다. 그리고, 상이성의 척도로서 BNNN과 Tanimoto, 2개의 방법들의 정확성을 비교하고자 한다. 이러한 방법들을 적용할 53203개의 화합물들의 효능과 화학적 구조가 있는 데이터를 이용한다.;It has been shown that the sequential design is efficient in large chemical databases with increasing numbers of compounds to be tested. Also, it is known that the multiple trees are more accurate than the greedy tree. We propose a sequential approach with multiple trees using the sequential design and the multiple trees. And, we compare the accuracy of the two methods, BNNN and Tanimoto for the measure of similarity. We use the data in the potency and chemical structure of 53203 compounds to apply those methods.-
dc.description.tableofcontentsCHAPTER 1 INTRODUCTI0N 1 CHAPTER 2 DATASET 3 CHAPTER 3 BACKGROUND 5 3.1 Recursive partitioning 5 3.2 Multiple Trees and Combining 7 CHAPTER 4 SEQUENTIAL ALGORITHM 8 4.1 Desingning a sequential screening scheme-Multiple Trees 8 Ⅰ) The initial sample size N1=2500 8 Ⅱ) Beam size b=5 ,pool size p=1000 8 Ⅲ) Additional sample size N2=2500, Design of additional stages D2=50/50 9 Ⅳ) Generating multiple trees(b=5,p=1000) 9 4.2 Designing a sequential screening scheme-Greedy Tree 10 Ⅰ) The initial sample size N1=2500 10 Ⅱ) Beam size b=5 ,pool size p=1000 10 Ⅲ) Additional sample size N2=2500, Design of additional stages D2=50/50 10 Ⅳ) Generating multiple trees(b=1,p=1) 11 4.3 BNNN and Tanimoto coefficient 12 CHAPTER 5 COMPARISON 13 5.1 Hit ratio and Gain ratio 13 5.2 Comparison 15 5.3 Relative hit ratio and Comparison 19 CHAPTER 6 DISCUSSION 24 REFERENCES 25 논문초록 26 감사의 글 27-
dc.formatapplication/pdf-
dc.format.extent808975 bytes-
dc.languageeng-
dc.publisher이화여자대학교 대학원-
dc.titleA sequential approach with multiple trees in large chemical databases-
dc.typeMaster's Thesis-
dc.format.page26 p.-
dc.identifier.thesisdegreeMaster-
dc.identifier.major대학원 통계학과-
dc.date.awarded2000. 8-
Appears in Collections:
일반대학원 > 통계학과 > Theses_Master
Files in This Item:
There are no files associated with this item.
Export
RIS (EndNote)
XLS (Excel)
XML


qrcode

BROWSE