发帖
雷达卡
华东师范大学统计系举办的2009暑期短课程讲义。主讲是威斯康星大学统计系教授邵军博士。
内容：
1. Introduction
2. Missing Mechanisms
(a) Missing completely at random
(b) Missing at random (ignorable missing)
(c) Covariate-dependent (unconfounded) missing
(d) Nonignorable missing
(e) Informative missing in longitudinal data
3. Ignorable Missing
(a) Parametric likelihood-based analysis under MAR
i. Bivariate normal data with one variable having
missing values
ii. Multivariate normal data with monotone missing
iii. EM algorithm
iv. Theory of the EM algorithm
v. Information and variance estimation
(b) Semi- and non-parametric methods under covariatedependent
missing
i. Linear models
ii. Nonparametric regression
iii. Re-weighting
iv. Empirical likelihood
v. Pseudo empirical likelihood
(c) Imputation methodology
i. Deterministic imputation
ii. Random imputation
(d) Variance estimation and inference
i. Multiple imputation
ii. Direction derivation with single imputation
iii. Resampling methods with single imputation
4. Nonignorable Missing
(a) Parametric likelihood approach
(b) Semi-parametric methods
(c) Empirical likelihood approach
(d) Pattern-mixture models
5. Longitudinal Data with Missing Values
(a) Informative missing
i. Parametric likelihood approach
ii. ACM (approximate conditional model)
(b) Grouping
(c) Pattern mixture
(d) Longitudinal imputation
i. Monotone missing
ii. Non-monotone missing
6. Dropout in Longitudinal Studies
(a) Treatment effect in a clinical trial
i. Study-end treatment effect (the traditional approach)
ii. Last-observed treatment effect (a new approach)
(b) Last observation carry forward (LOCF)
(c) Last observation analysis
7. Others
(a) Measurement error and missing data
(b) Missing data in covariates
(c) Omitted covariates
8. Concluding Remarks
京公网安备 11010802022788号
