arXiv：cs.LG· Gabriel Nova, Stephane Hess, Sander van Cranenburgh·· 3 小时前
Delphos：用深度强化学习辅助离散选择模型设定的框架
Delphos: A reinforcement learning framework for assisting discrete choice model specification
AI 导读
Delphos 是一个深度强化学习框架，用于辅助离散选择模型的设定过程，通过自动化的数据驱动建议减少开发者构建和优化效用函数的工作量。该框架将模型设定建模为序列决策问题，使用 Deep Q-Network 学习添加替代特定常数、设定口味参数、非线性变换及协变量交互等建模动作的效果。在模拟和实证数据集上，Delphos 仅探索少量可行建模空间便学得有效设定策略，并能生成具有竞争力的行为合理模型。
正文
Abstract:We introduce Delphos, a deep reinforcement learning framework for assisting discrete choice model specification process. Delphos aims to support the modeller by providing automated, data-driven suggestions for model specifications, thereby reducing the effort required to develop and refine utility functions. Delphos conceptualises model specification as a sequential decision-making problem, inspired by the way human choice modellers iteratively construct models through a series of reasoned specification decisions. In this setting, an agent learns to specify candidate model specifications by choosing a sequence of modelling actions, such as adding alternative specific constants, accommodating both generic and alternative-specific taste parameters, applying non-linear transformations to attributes, and including interactions with covariates. Each resulting candidate model is estimated and evaluated using a reward function defined by the modeller, which can reflect statistical model fit as well as behavioural expectations. Specifically, Delphos uses a Deep Q-Network to learn how individual specification decisions contribute to the eventual quality of the resulting model and, in turn, which sequences of modelling decisions tend to produce well-performing candidates. We evaluate Delphos on both simulated and empirical datasets using alternative reward functions. In simulated cases, learning curves, Q-value patterns, and performance metrics show that Delphos learns effective specification strategies while exploring only a small fraction of the feasible modelling space. We further apply the framework to two empirical datasets to benchmark and demonstrate its practical use. These experiments illustrate the ability of Delphos to generate competitive, behaviourally plausible models and highlight the potential of this adaptive, learning-based framework to assist the model specification process.
| Comments: | 13 pages, 7 figures |
| Subjects: | General Economics (econ.GN); Machine Learning (cs.LG) |
| Cite as: | arXiv:2506.06410 [econ.GN] |
| (or arXiv:2506.06410v4 [econ.GN] for this version) | |
| https://doi.org/10.48550/arXiv.2506.06410 arXiv-issued DOI via DataCite |
Submission history
From: Gabriel Nova [view email]
[v1]
Fri, 6 Jun 2025 15:40:16 UTC (2,224 KB)
[v2]
Fri, 25 Jul 2025 13:23:22 UTC (2,827 KB)
[v3]
Mon, 16 Mar 2026 10:58:11 UTC (1,977 KB)
[v4]
Thu, 8 Oct 2026 08:02:22 UTC (1,892 KB)
来源：arXiv：cs.LG · arxiv.org