Concordance-based survival analysis and clustering
Grant Data
Project Title
Concordance-based survival analysis and clustering
Principal Investigator
Professor Yin, Guosheng
(Principal Investigator (PI))
Duration
36
Start Date
2021-10-01
Amount
512503
Conference Title
Concordance-based survival analysis and clustering
Presentation Title
Keywords
Concordance index, Kaplan-Meier estimator, Network data, Recurrent events, Survival function
Discipline
Probability & Statistics,Clinical Trials
Panel
Physical Sciences (P)
HKU Project Code
17308321
Grant Type
General Research Fund (GRF)
Funding Year
2021
Status
On-going
Objectives
1 We propose a new general class of nonparametric survival functions using the Box-Cox transformation, which include the Kaplan-Meier estimator as a special case. Such a general class provides a more flexible fit to the survival function by assigning different weights to different follow-up periods. 2 We develop concordance-based regression with recurrent event data and quantify the concordance between progression-free survival and overall survival to provide guidance on when the former can be used as a surrogate for the later. 3 For social network data, we propose a new clustering algorithm based on triangular concordance which can account for both the heterogeneity and transitivity within clusters.