Fully Data-driven but Interpretable Human Behavioural Modelling with Differentiable Discrete Choice Model

Fumiyasu Makinoshima, Tatsuya Mitomi, Fumiya Makihara, Eigo Segawa · 2024

Read Fully Data-driven but Interpretable Human Behavioural Modelling with Differentiable Discrete Choice Model by Fumiyasu Makinoshima, Tatsuya Mitomi, Fumiya Makihara, Eigo Segawa (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-19403.txt.

Description

Fumiyasu Makinoshima, Tatsuya Mitomi, Fumiya Makihara, and Eigo Segawa present a differentiable discrete choice model in two thousand twenty-four. Choice models track how humans pick among options. Their setups have long leaned on expert domain knowledge. Full auto modeling that stays readable has been a hard goal.

Their model learns from features and choices alone, with no prior facts. It yields utility functions in closed form that can be read. Tests on fake and real data span many data types. Runs finish in tens of seconds on a plain laptop. The model's smooth form also maps paths to shift behavior.

Catalog record

CREATORFumiyasu Makinoshima, Tatsuya Mitomi, Fumiya Makihara, Eigo Segawa
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE38 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-19403

Browse the full library · About the Tower of Babel