Causal Inference on Time Series using Restricted Structural Equation Models

By Jan Peters, Dominik Janzing, Bernhard Schölkopf. 2013

209 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Telling cause from effect in deterministic linear dynamical systems (2015, background): “In particular, this is the case when unknown time lags are introduced in the measured time series.”
Automated Machine Learning: Methods, Systems, Challenges (2019, unknown): “Reference [26], for instance, shows that there exist statistical dependences between two quantum systems A and B that uniquely indicate whether one is the cause of the other or whether there is a common cause of both, see Fig.”
Identifiability and estimation of structural vector autoregressive models for subsampled and mixed-frequency time series. (2017, methodology): “Our work concretely connects the non-Gaussian subsampled VAR with independent innovations method [14] with the now extensive non-Gaussian SVAR framework [15, 16, 17, 18, 19] by proving identifiability of an SVAR model of order one under arbitrary subsampling.”
A Survey of Learning Causality with Data: Problems and Methods (2018, methodology): “[105] proposed a more robust algorithm based on non-linear independent tests, known as TiMINo, which can avoid discovering incorrect causal relationships under model misspecification.”
Causal Discovery with Attention-Based Convolutional Neural Networks (2019, methodology): “(The authors present two other variants besides the linear model, of which ‘TiMINo-GP’ was shown to be more suitable for time series with more than 300 time steps [22], but only the linear model was fully implemented by the authors.)”
Masked Gradient-Based Causal Structure Learning (2019, methodology): “For time series data, several methods [27, 24, 6] ﬁrst choose a time lag p ≥ 0 and then perform causal discovery among d ( p + 1) variables, where many edges can be excluded as the variable at the current time cannot be a cause of the one of past time.”
Identification of Time-Dependent Causal Model: A Gaussian Process Treatment (2015, background): “In particular, [Peters et al., 2013] considered the nonlinear causal relations,”
Discovering Temporal Causal Relations from Subsampled Data (2015, methodology): “We conducted experiments on the Temperature Ozone data and the Temperature in House data (Peters et al., 2013).”
A Survey of Learning Causality with Data (2018, methodology): “lass of algorithms for learning causal relationship in time series are based on the statistical independence tests. Researchers [29, 39] proposed to adapt FCI algorithm for time series. Peters et al. [105] proposed a more robust algorithm based on non-linear independent tests, known as TiMINo, which”