Causal inference for time series analysis: problems, methods and evaluation

Causal discovery of drivers of surface ozone variability in Antarctica using a deep learning algorithm

“…The time series X is said to be a Granger cause of Y , if the past values of X provide information about the current state of Y , beyond what is already known from the past values of Y alone. Mathematically, linear VAR models are represented as follows: 26 where vector Y t = [ Y 1 t ,…, Y kt ] contains the values of k at time step t ; ψ ( T ) is the ( k × k ) coefficient matrix at the time lag T and ε t is noise term (residual) assumed to be independent and identically distributed. Here, T max is the maximum time lag.…”

Causal Discovery in Manufacturing: A Structured Literature Review

“…Notably, initial successful applications have been noted in various fields of research, such as medicine [15], genetics [16], natural sciences [17], ecology [18], astronomy [19], and neuroscience [20]. Due to these developments, several literature reviews can be found on causal discovery method concepts and benchmarks [21,22], as well as their applications in earth system sciences [23] or biomedical informatics [24]. However, due to their specific nature and progression complexity, the insights from these reviews cannot be transferred to manufacturing.…”

“…These are constraint-based methods, scorebased methods, hybrid models, and methods based on functional causal models [21,38]. Furthermore, in the context of time-series data, several other methods can be identified, particularly methods based on Granger causality and conditional independence, as well as methods based on structural equation models and deep learning [22].…”

Evaluation Methods and Measures for Causal Learning Algorithms

“…The author especially examines where the links between AI and graphical causal inference have been and should be established. The rapid growth of time-series data and the unique challenges it brings in causal studies have led to surveys such as [24] reviewing problems, methods, and evaluation related to causal time series analysis. In causal ML, Chen et al [25] summarize different types of bias in recommendation systems and review methods that aim to mitigate such bias by leveraging causal inference theories.…”