arxivst stuff from arxiv that you should probably bookmark

Treatment-Response Models for Counterfactual Reasoning with Continuous-time, Continuous-valued Interventions

Abstract · Apr 6, 2017 22:42 ·

markers treatment response 21218 creatinine baltimore interventions hopkins johns stat-ml cs-ai cs-lg

Arxiv Abstract

  • Hossein Soleimani
  • Adarsh Subbaswamy
  • Suchi Saria

Treatment effects can be estimated from observational data as the difference in potential outcomes. In this paper, we address the challenge of estimating the potential outcome when treatment-dose levels can vary continuously over time. Further, the outcome variable may not be measured at a regular frequency. Our proposed solution represents the treatment response curves using linear time-invariant dynamical systems—this provides a flexible means for modeling response over time to highly variable dose curves. Moreover, for multivariate data, the proposed method: uncovers shared structure in treatment response and the baseline across multiple markers; and, flexibly models challenging correlation structure both across and within signals over time. For this, we build upon the framework of multiple-output Gaussian Processes. On simulated and a challenging clinical dataset, we show significant gains in accuracy over state-of-the-art models.

Read the paper (pdf) »