EM algorithms for nonparametric estimation of mixing distributions

By Kenneth Train. 2008

314 citations recorded by the source corpus.

Abstract This paper describes and implements three computationally attractive procedures for nonparametric estimation of mixing distributions in discrete choice models. The procedures are specific types of the well known EM (Expectation-Maximization) algorithm based on three different ways of approximating the mixing distribution nonparametrically: (1) a discrete distribution with mass points and frequencies treated as parameters, (2) a discrete mixture of continuous distributions, with the moments and weight for each distribution treated as parameters, and (3) a discrete distribution with fixed mass points whose frequencies are treated as parameters. The methods are illustrated with a mixed logit model of households' choices among alternative-fueled vehicles.

Subfields: Automotive Engineering, Economics and Econometrics, Renewable Energy, Sustainability and the Environment

View Paper: https://www.semanticscholar.org/paper/5ca8cef36e7380a9f42d4055ab33c81ecee36d6e

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Automotive Engineering

In Economics and Econometrics

In Renewable Energy, Sustainability and the Environment

Citations in the corpus

Latent class structures: taste heterogeneity and beyond (2014, methodology): “…with larger numbers of classes can be problematic with parameters collapsing to the same values across classes or some classes obtaining very small probabilities, and here, the EM algorithm can once again be one possible solution, discussed in Train (2008) but also earlier on by Bhat (1997).”
Assuring finite moments for willingness to pay in random coefficient models (2009, unknown): “• Another possible solution is the use of a finite mixture models, including latent class models (see e.g. Hess et al. 2007) and the non-parametric estimation procedures suggested by Train (2008).”
COMPARING ALTERNATIVE MODELS OF HETEROGENEITY IN CONSUMER CHOICE BEHAVIOR (2012, unknown): “Train (2008) and Bajari et al (2007) also consider specifying mixture-of-normals for MIXL but both studies focus on alternative algorithms rather than the performance of MM-MNL compared to existing models.”
A stated choice experiment to measure the effect of informational and normative conformity in the preference for electric vehicles (2017, result): “This interpretation is in line with the negative effect of the “additional detour time due to limited charging/refuelling locations” that was found in Train (2008) and Hoen and Koetse (2012).”
The random coefficients logit model is identified (2012, methodology): “…of heterogeneous parameters β. 7As a referee points out, the sign restriction or the monotonicity assumption for a special regressor may not be too restrictive because it is testable and the sign can be determined from a pre-model analysis as described in Lewbel (2000) and his other related…”
Are consumers willing to pay to let cars drive for them? Analyzing response to autonomous vehicles (2017, background): “, Q} or fΘ(θi) = (5)Any continuous distribution can be approximated by a discrete mixture of normal distributions (Train, 2008).”