Discussion of paper by K. Garside et al.

Moo K. Chung, Hernando Ombao

Research output: Contribution to journalArticlepeer-review


We discuss the paper Garside et al. (2020). Although topological data analysis (TDA) has been around for many decades with well grounded theoretical development, it still suffers from numerous statistical and computational issues. For these reasons, it has not yet become a standard tool for data scientists. The authors point out the difficulty of directly applying existing statistical models to persistent homology due to the heterogeneous nature of topological features. The statistical development in TDA in the last decade has been focused on making heterogeneous features into homogenous structured data by transformations or smoothing. The main achievement is that stability results on such transformed topological features have been established. Thus, the idea of applying survival analysis techniques to the birth and death process of topological feature is very intriguing. The authors succeed in elucidating the connection between the event history methods and the life time of topological features. However, the paper has stimulated new interesting questions.
Original languageEnglish (US)
JournalAccepted by Biometrika
StatePublished - 2021

Bibliographical note

KAUST Repository Item: Exported on 2021-05-05
Acknowledgements: We would like to thank Robin Henderson of Newcastle University, U.K. for providing the coordinates and connectivity information of nodes in the tree used in displaying Figure 1. This study is supported by NIH grants R01 EB022856 and R01 EB028753, NSF grant MDS-2010778,
and CRG from the King Abdullah University of Science and Technology.


Dive into the research topics of 'Discussion of paper by K. Garside et al.'. Together they form a unique fingerprint.

Cite this