arXiv：cs.LG· Carlos Heredia, Daniel Roncel·· 2 天前AI 评分28
ICDN：基于神经需求势函数的可积弹性模型
Integrable Elasticity via Neural Demand Potentials
AI 导读
研究者提出 ICDN（Integrable Context-Dependent Demand Network），一种面向多产品零售需求的"需求优先"神经模型，将对数需求学习为对数价格的平滑、上下文条件函数，从而可从学到的需求曲面精确推导弹性。在三个零售数据集上，该模型取得有竞争力的样本外预测表现，并给出解析可处理、经济上正则化的自身与交叉价格响应。
正文
Abstract:We propose the Integrable Context-Dependent Demand Network (ICDN), a demand-first neural model for multiproduct retail demand. ICDN learns log-demand as a smooth, context-conditioned function of log-prices, allowing elasticities to be derived exactly from the learned demand surface. Across three retail datasets, the model achieves competitive out-of-sample prediction while producing analytically tractable, and economically regularized own- and cross-price responses.
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2605.22820 [cs.LG] |
| (or arXiv:2605.22820v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2605.22820 arXiv-issued DOI via DataCite |
|
| Related DOI: | https://doi.org/10.1016/j.eswa.2026.134600
DOI(s) linking to related resources |
Submission history
From: Carlos Heredia Pimienta [view email]
[v1]
Thu, 21 May 2026 17:59:47 UTC (857 KB)
[v2]
Thu, 1 Oct 2026 08:04:58 UTC (719 KB)
来源：arXiv：cs.LG · arxiv.org