Shigeki Kishi

This paper introduces FORCC, a causal inference framework for ecological time series, which identifies directional causality and reduces false positives from synchronization, revealing ecological interactions in marine data.

This paper introduces F-CODEC, a causal discovery method that reduces false positives in nonlinear dynamical systems without surrogate data, validated on ecological time series.

This paper introduces FORWS and FORWC methods for causal inference in high-dimensional time series, showing improved forecasting and noise resilience using forest weights instead of Euclidean metrics.