A scoping methodological review of simulation studies comparing statistical and machine learning approaches to risk prediction for time-to-event data

Table 5 Failure time distributions, assumptions and covariate effects included in the data-generating mechanisms for each article

*Geng et al. (2014) [26] included a specific crossing hazards data-generating mechanism
**Gong et al. (2018) [27] take the exponential of the first covariate squared and cos transform second covariate; covariate coefficients were also obtained for the clinically relevant data-generating mechanisms by fitting each of the predefined models to clinical data
***Katzman et al. (2018) [29] use a Gaussian distribution for the linear predictor and include quadratic effects for both covariates
****Lowsky et al. (2012) [25] fit an exponential model to the clinical data to obtain estimates for the covariate coefficients to use in simulating the failure times
*****Wang and Li (2019) [30] transform the covariates by a radial basis kernel

ISSN: 2397-7523