论文信息 - A NOVEL NEURAL NETWORK FOR SURVIVAL ANALYSIS

A NOVEL NEURAL NETWORK FOR SURVIVAL ANALYSIS

In this paper a feedforward neural network architecture to model survival probabilities is illustrated. The two main features of the network are that non-linearity are captured in the survival function, and the time variable is embedded in the model so it is able to extract its interactions with other system features. The model is described in a hierarchical Bayesian framework. Some experiments with synthetic and real world data show the capabilities of the model.

Roberto Tagliaferri | L. Milano | A. Eleuteri | Sabino De Placido | M. De Laurentiis

[1] D. Rubin,et al. Inference from Iterative Simulation Using Multiple Sequences , 1992 .

[2] W. Linde. STABLE NON‐GAUSSIAN RANDOM PROCESSES: STOCHASTIC MODELS WITH INFINITE VARIANCE , 1996 .

[3] Laurence L. George,et al. The Statistical Analysis of Failure Time Data , 2003, Technometrics.

[4] W. Vach,et al. On the misuses of arti"cial neural networks for prognostic and diagnostic classi"cation in oncology , 2000 .

[5] Richard L. Tweedie,et al. Markov Chains and Stochastic Stability , 1993, Communications and Control Engineering Series.

[6] Geoffrey E. Hinton,et al. Bayesian Learning for Neural Networks , 1995 .

[7] Radford M. Neal. Slice Sampling , 2003, The Annals of Statistics.

[8] Heekuck Oh,et al. Neural Networks for Pattern Recognition , 1993, Adv. Comput..

[9] Adrian E. Raftery,et al. Accounting for Model Uncertainty in Survival Analysis Improves Predictive Performance , 1995 .