名称：Exposure, hazard, and survival analysis of diffusion on social networks
published in statistics in medicine.
描述：Social network is important for the diffusion of ideas and products through human societies. People will tend to influence their friends on their decision to use certain products. We are interested in using the network information to maximize the adoption of products on social networks. Due to the resource constraints, the questions lie in which individuals should we target initially so that more people can use the products. My collaborators conducted randomized clinical trials in Honduras to answer this question for the diffusion of multivitamin tablets. We use a survival analysis framework to analyze this type of data. We find that targeting the most influential people in the network will lead to the maximum adoption of the multivitamin tablets. This conclusion will help future researchers to better allocate resources for the adoption of products to help people in developing countries
名称：Flexible and interpretable models for survival data
描述：As data sets continue to increase in size, there is growing interest in methods for prediction that are both flexible and interpretable. A flurry of recent work on this topic has focused on additive modeling in the regression setting, and in particular, on the use of data-adaptive non-linear functions that can be used to flexibly model each covariate's effect, conditional on the other features in the model. In this paper, we extend this recent line of work to the survival setting. We develop an additive Cox proportional hazards model, in which each additive function is obtained by trend filtering, so that the fitted functions are piece-wise polynomial with adaptively-chosen knots. An efficient proximal gradient descent algorithm is used to fit the model. We demonstrate its performance in simulations and in application to a primary biliary cirrhosis (PBC) dataset.
- Develop penalized regression (machine learning) models to fit piecewise polynomial in Cox proportional hazard model.
- Model covariate effects flexibly using trend filtering penalty.
- Knots are chosen adaptively based on the data.