A distribution-to-distribution regression model, validated on London Underground data, predicts station exit-count distributions under novel disruptions as learned linear combinations of natural-regime feature distributions.
Predicting the future behavior of a time-varying probability distribution
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Counterfactual Distribution Regression for Structured Inference
A distribution-to-distribution regression model, validated on London Underground data, predicts station exit-count distributions under novel disruptions as learned linear combinations of natural-regime feature distributions.