Machine learning and cox based benchmarking tool: exploration of survival models associated with chronic degenerative diseases

the present work reports the exploration of the CoxBenchmarking function applied to chronic-degenerative disease datasets associated with survival. CoxBenchmarking implementation is a computer-based benchmarking algorithm that compares the Survival Models that were constructed by several machine learning strategies. It was developed as an extension of FRESA.CAD package and uses its Random Holdout Cross-Validation. CoxBenchmarking provides an algorithm that generates eleven distinct survival models through feature selection of ML-based techniques: 6 wrappers and 5 filters. Besides, the function summarizes the results with tables and graphs by providing a well-ordered data structure and a plot function.

https://orcid.org/0000-0003-1361-5162