Generating Survival Interpretable Trajectories and Data
This work addresses the need for interpretable survival analysis tools in domains like healthcare, offering incremental improvements through a novel autoencoder structure and weighting scheme.
The authors tackled the problem of generating interpretable survival trajectories and data by proposing a new autoencoder-based model that predicts event times, generates supplementary data, and produces prototype trajectories for counterfactual explanations. The model demonstrated efficiency in numerical experiments on synthetic and real datasets, with code made publicly available.
A new model for generating survival trajectories and data based on applying an autoencoder of a specific structure is proposed. It solves three tasks. First, it provides predictions in the form of the expected event time and the survival function for a new generated feature vector on the basis of the Beran estimator. Second, the model generates additional data based on a given training set that would supplement the original dataset. Third, the most important, it generates a prototype time-dependent trajectory for an object, which characterizes how features of the object could be changed to achieve a different time to an event. The trajectory can be viewed as a type of the counterfactual explanation. The proposed model is robust during training and inference due to a specific weighting scheme incorporating into the variational autoencoder. The model also determines the censored indicators of new generated data by solving a classification task. The paper demonstrates the efficiency and properties of the proposed model using numerical experiments on synthetic and real datasets. The code of the algorithm implementing the proposed model is publicly available.