Causal Discovery on Irregular Time Series
2026-07-20 • Machine Learning
Machine Learning
AI summaryⓘ
The authors address the challenge of finding cause-and-effect relationships in data that is collected at irregular times, which is common in areas like healthcare and finance. They extend an existing method called PCMCI+ that usually works only with regularly spaced data by allowing it to look at influence over flexible time windows instead of fixed delays. Their tests on made-up irregular data show that their new approach does a better job at discovering true causal links compared to the original method. This means their method is more useful for real-world situations where data isn't evenly spaced.
causal discoverytime seriesirregular samplingPCMCI+causal graphsignal-to-noise ratiomultivariate datatemporal windowssynthetic datacausal influence
Authors
Martim Penim, Ricardo Ribeiro Pereira, Jacopo Bono, Hugo Ferreira, Mário A. T. Figueiredo, Pedro Bizarro
Abstract
Causal discovery methods have shown strong performance in temporal systems, but they typically rely on regular and discrete lag structures, limiting their applicability to regularly sampled data. However, many real-world tasks require dealing with irregularly sampled streams of events, such as sensor streams, healthcare data, and financial transactions. In this work, we propose an extension of PCMCI+, a state-of-the-art method for causal discovery on regular multivariate time series, to allow for handling irregular time series. Instead of modelling causal relations through fixed-lag dependencies, our method aggregates causal influence over predefined temporal windows. We evaluate our method on synthetic irregular event streams with known causal structures under different signal-to-noise ratios, showing that it consistently recovers the underlying causal graph and substantially outperforms the standard PCMCI+ on irregularly sampled data.