Causal Discovery from Subsampled Time Series Data by Constraint Optimization
This addresses a practical issue in fields like neuroscience or economics where data collection limitations lead to subsampling, though it appears incremental as it builds on prior work on subsampling effects.
The paper tackles the problem of estimating causal structure from time series data when measurements are collected at a coarser timescale than the underlying causal processes, which can cause significant errors. They propose a constraint optimization approach that optimally handles statistical conflicts and achieves computational performance several orders of magnitude better than previous methods.
This paper focuses on causal structure estimation from time series data in which measurements are obtained at a coarser timescale than the causal timescale of the underlying system. Previous work has shown that such subsampling can lead to significant errors about the system's causal structure if not properly taken into account. In this paper, we first consider the search for the system timescale causal structures that correspond to a given measurement timescale structure. We provide a constraint satisfaction procedure whose computational performance is several orders of magnitude better than previous approaches. We then consider finite-sample data as input, and propose the first constraint optimization approach for recovering the system timescale causal structure. This algorithm optimally recovers from possible conflicts due to statistical errors. More generally, these advances allow for a robust and non-parametric estimation of system timescale causal structures from subsampled time series data.