Survival Analysis with Multivariate Adaptive Regression Splines using Cox-Snell Residual

No Thumbnail Available



Journal Title

Journal ISSN

Volume Title


Faculty of Computer and Applied Computer Science, Tibiscus University of Timisoara, Romania


Multivariate Adaptive Regression Splines (MARS) are a generalization of stepwise linear regression method that is often employed to improve the efficiency of regression models. It is a useful tool to identify linear/nonlinear and interactions effects between a set of metrical and categorical covariates in regression models. In this study, the use of a modified Cox-Snell Residuals to Survival Analysis with MARS was proposed. The proposed method was compared with Martingale Residual in the Survival MARS setting. These two residual types were used as responses in the Cox proportional hazard modeling in the MARS implementations. Results from simulation studies revealed that the proposed method fitted the data better than the Martingale residual However, further results from Monte-Carlo experiment showed that the two residual types performed better than the classical Cox Proportional Hazard (CPH) method. These methods were applied on real life dataset on Pneumocystis Carinii Pneumonia and all the results obtained actually validated those got from the simulation studies



MARS, Martingale Residual, Modified Cox-Snell Residual, PCP, CART