We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Dynamic Discrete Choice Estimation with Partially Observable States and Hidden Dynamics

Abstract: Dynamic discrete choice models are used to estimate the intertemporal preferences of an agent as described by a reward function based upon observable histories of states and implemented actions. However, in many applications, such as reliability and healthcare, the system state is only partially observable or hidden (e.g., the level of deterioration of an engine, the condition of a disease), and the decision maker only has access to information imperfectly correlated with the true value of the hidden state. In this paper, we consider the estimation of a dynamic discrete choice model with state variables and system dynamics hidden to both the agent and the modeler, thus generalizing the model in Rust(1987) to partially observable cases. We examine the structural properties of the model and prove that this model is still identifiable if the cardinality of the state space, the discount factor, the distribution of random shocks, and the rewards for a given (reference) action are given. We analyze both theoretically and numerically the potential mis-specification errors that may be incurred when the Rust's model is improperly used in partially observable settings. We further apply the model to a subset of dataset in Rust(1987) for bus engine mileage and replacement decisions. The results show that our model can improve model fit as measured by the $\log$-likelihood function by $17.73\%$ and the $\log$-likelihood ratio test shows that our model statistically outperforms the Rust's model. Interestingly, our hidden state model also reveals an economically meaningful route assignment behavior in the dataset which was hitherto ignored, i.e. routes with lower mileage are assigned to buses believed to be in worse condition.
Subjects: Machine Learning (cs.LG); Machine Learning (stat.ML)
Cite as: arXiv:2008.00500 [cs.LG]
  (or arXiv:2008.00500v2 [cs.LG] for this version)

Submission history

From: Zhide Wang [view email]
[v1] Sun, 2 Aug 2020 15:04:27 GMT (295kb,D)
[v2] Fri, 27 Nov 2020 20:22:07 GMT (226kb,D)

Link back to: arXiv, form interface, contact.