A1380
Title: On survival trees for interval-censored survival data
Authors: Asanao Shimokawa - Tokyo University of Science (Japan) [presenting]
Abstract: Generally, survival time analysis often considers only right-censored data. In the construction of survival trees, many studies also assume right-censoring. However, in practice, many data sets are more appropriately modeled under the assumption of interval-censoring rather than right-censoring. Accordingly, the focus is on the construction of survival trees that account for interval censoring. For this purpose, approaches are proposed based on hypothesis tests as well as on prediction accuracy metrics. The performance of this criterion is evaluated through simulation studies. In addition to this, the results of the survival tree obtained from actual data are shown.
