A study on exploratory identification of responsive subgroups in complex survival event data

2017 Fiscal Year Final Research Report

• Project/Area Number: 15K15950
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Statistical science
• Research Institution: Keio University
• Principal Investigator: Hayashi Kenichi 慶應義塾大学, 理工学部(矢上), 講師 (70617274)
• Project Period (FY): 2015-04-01 – 2018-03-31
• Keywords: 医学統計学 / 区間打ち切りデータ / C-index / IDI / NRI / ROC解析 / power-IDI / クラスタワイズ回帰モデル

Outline of Final Research Achievements

This project aims to develop measures of statistical model for survival outcomes and related variables such as biomarkers, and methods that detects responsive subgroups in data. As for the first issue, an estimation method of C-index for interval censored data is proposed. Further, a modification of the IDI (integrated discriminant improvement) is developed to overcome critical concerns in the original IDI. As for the second issue, a cluster-wise linear model is extended based on the generalized mean, which is called Kolmogorov-Nagumo average. The extended model involves several models proposed in the previous study as special caes. Through numerical experiments, the superiority of the proposed model is implied.

Free Research Field

統計科学