A binary-code experiment design with O(log n) assortments, plus a boost-factor algorithm, provably recovers substitution nests in Nested Logit models and improves choice prediction.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ME 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Experimental Assortments for Choice Estimation and Nest Identification
A binary-code experiment design with O(log n) assortments, plus a boost-factor algorithm, provably recovers substitution nests in Nested Logit models and improves choice prediction.