View : 740 Download: 0

Full metadata record

DC Field Value Language
dc.contributor.advisor이은경-
dc.contributor.author고명지-
dc.creator고명지-
dc.date.accessioned2020-02-03T16:33:07Z-
dc.date.available2020-02-03T16:33:07Z-
dc.date.issued2020-
dc.identifier.otherOAK-000000163372-
dc.identifier.urihttp://dcollection.ewha.ac.kr/common/orgView/000000163372en_US
dc.identifier.urihttps://dspace.ewha.ac.kr/handle/2015.oak/253245-
dc.description.abstractThe visual predictive check (VPC) method is a popular way to validate models in pharmacometric area. However, the binning of the independent variable makes it less effective, and the precision of a VPC plot decreases, especially when the data are observed sparsely. To improve the precision of the VPC, we adapted the idea of the average shifted histogram (ASH), and propose the average shifted visual predictive check (asVPC). asVPC aggregates the information from several neighboring bins. With this approach, we can overcome current problems with VPC, especially from data sparsity. Also, it allows for more precise decisions for model diagnostics. The asVPC produces clearer and more precise plots than the original VPC method.;약동학에서 사용되는 모델의 진단방법 중 VPC(Visual Predictive Check)가 가장 널리 쓰인다. 하지만 얻어진 관측치가 넓게 퍼져 있으면 독립 변수에 따라 범례를 나누는 작업이 VPC의 정확도가 떨어지게 할 수 있다. 본 논문에서는 이를 해결하기 위해 해당 범례의 정보와 주변에 있는 범례의 정보에 가중치를 두어 결합을 하는 ASH(Average Shifted Histogram)의 아이디어를 차용한 asVPC(Average Shifted Visual Predictive Check) 방법을 제안한다. asVPC는 이동 히스토그램의 개수인 𝑚에 따라 VPC의 신뢰구간 영역을 줄여 VPC의 정확도를 높일 수 있다.-
dc.description.tableofcontentsI. Introduction 1 II. Pharmacokinetic model and Model Diagnosis 3 A. Population model 3 B. Parameter estimation 4 C. Model Diagnosis 4 1. Residual plot 5 2. Normalized prediction distribution error 5 3. Visual Predictive Check 6 D. Binning 11 E. Additive quantile regression 12 III. Average Shifted Visual Predictive Check 13 A. Average Shifted Histogram 13 B. Algorithm of asVPC 14 1. Combining results from each bin 14 2. Deciding on the weights 15 IV. Simulation and Results 16 A. Simulation 17 1. Same model for original and simulated data 17 2. Different model for original and simulated data 20 B. Application 23 V. Discussion 26 Bibliography 27 Abstract (in Korean) 29-
dc.formatapplication/pdf-
dc.format.extent995695 bytes-
dc.languageeng-
dc.publisher이화여자대학교 대학원-
dc.subject.ddc500-
dc.titleNew Visual Diagnostic Tools for Pharmacometric Models-
dc.typeMaster's Thesis-
dc.format.pageiii, 29 p.-
dc.identifier.thesisdegreeMaster-
dc.identifier.major대학원 통계학과-
dc.date.awarded2020. 2-
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