While deep neural networks (DNNs) have been increasingly applied to choice analysis showing high predictive power, it is unclear to what extent researchers can interpret economic information from DNNs. This paper demonstrates that DNNs can provide economic information as complete as classical discrete choice models (DCMs). The economic information from DNNs includes choice predictions, choice probabilities, market shares, substitution patterns of alternatives, social welfare, probability derivatives, elasticities, marginal rates of substitution, and heterogeneous values of time. Unlike DCMs, DNNs can automatically learn utility functions and reveal behavioral patterns that are not prespecified by domain experts, particularly when the sample size is large. However, the economic information obtained from DNNs can be unreliable when the sample size is small, because of three challenges associated with the automatic learning capacity: high sensitivity to hyperparameters, model non-identification, and local irregularity. The first challenge is related to the statistical challenge of balancing approximation and estimation errors of DNNs, the second to the optimization challenge of identifying the global optimum in the DNN training, and the third to the robustness challenge of mitigating locally irregular patterns of estimated functions. To demonstrate the strength and challenges, we estimated the DNNs using a stated preference survey from Singapore and a revealed preference data from London, extracted the full list of economic information from the DNNs, and compared them with those from the DCMs. We found that the economic information either aggregated over trainings or population is more reliable than the disaggregate information of the individual observations or trainings, and that larger sample size, hyperparameter searching, model ensemble, and effective regularization can significantly improve the reliability of the economic information extracted from the DNNs. Future studies should investigate the requirement of sample size, better ensemble mechanisms, other regularizations and DNN architectures, better optimization algorithms, and robust DNN training methods to address DNNs three challenges to provide more reliable economic information for DNN-based choice models.

尽管深度神经网络（DNN）已越来越多地应用于具有高预测能力的选择分析，但尚不清楚研究人员在多大程度上可以解释DNN的经济信息。本文证明了DNN可以提供完整的经济信息作为经典的离散选择模型（DCM）。来自DNN的经济信息包括选择预测，选择概率，市场份额，替代品的替代模式，社会福利，概率导数，弹性，替代边际率和时间的异质性。与DCM不同，DNN可以自动学习实用功能并显示领域专家未指定的行为模式，尤其是在样本量较大时。但是，当样本量较小时，从DNN获得的经济信息可能不可靠，这是因为与自动学习能力相关的三个挑战：对超参数的高灵敏度，模型无法识别以及局部不规则。第一个挑战与平衡DNN逼近和估计误差的统计挑战有关，第二个挑战与DNN训练中确定全局最优值的优化挑战有关，第三个与缓解估计函数的局部不规则模式的鲁棒性挑战有关。为了展示其优势和挑战，我们使用新加坡的既定偏好调查和伦敦公开的偏好数据估算了DNN，从DNN中提取了完整的经济信息清单，并将其与DCM进行了比较。我们发现，在培训或总体上汇总的经济信息比单个观察或培训的分类信息更可靠，并且样本量更大，超参数搜索，模型集成，有效的正则化可以显着提高从DNN中提取的经济信息的可靠性。未来的研究应该调查样本量，更好的集成机制，其他正则化和DNN架构，更好的优化算法以及健壮的DNN训练方法的需求，以解决DNN的三个挑战，从而为基于DNN的选择模型提供更可靠的经济信息。