Matches in SemOpenAlex for { <https://semopenalex.org/work/W2933596316> ?p ?o ?g. }
Showing items 1 to 61 of
61
with 100 items per page.
- W2933596316 abstract "The objectives of this dissertation are three-fold: (1) to compare bivariate normal approximation algorithms in terms of accuracy and efficiency; (2) to develop a more flexible maximum likelihood algorithm for the trichotomous probit model; and (3) to evaluate the simple logit, nested logit, independent probit, and trichotomous probit models with regard to theoretical flexibility, estimation and predictive accuracy, and computational ease using simulated data.The comparative analysis of approximation algorithms is performed to improve upon the computational burden of existing probit algorithms. Of the eight methods investigated, the method of Divgi (1979) emerged as the fastest and most accurate.Flexibility is achieved in two ways. First, the reduced-form trinomial probit model developed by Terza (1985a), which allows for full covariance flexibility, is extended to encompass a wider range of applications. Second, a software package offering the user various options to be invoked interactively is coded in FORTRAN 77.The performance of the four models is evaluated in terms of the bias and mean square error (MSE) of the elasticities of the choice probabilities with respect to the regressors (evaluated at regressor means), the predictive accuracy, and the number of iterations required for convergence. The results can be summarized as follows. First, the trichotomous probit model proved to be best with respect to the bias and MSE of the elasticity estimates. Second, predictive accuracy did not provide a means for discriminating among the models. However, with a likelihood ratio test and the t-tests, we rejected the simple logit and independent probit null hypotheses, a result that is consistent with the data generation process. Moreover, the logit models required fewer iterations for convergence.Finally, the bias and MSE of the parameter estimates for the trichotomous probit model are presented. The results support the contention that probit-type choice models should not be rejected for reasons of computational impracticality. Moreover, single-precision for the approximation of the bivariate normal distribution gives sufficient accuracy and does not increase the number of iterations required." @default.
- W2933596316 created "2019-04-11" @default.
- W2933596316 creator A5014766715 @default.
- W2933596316 creator A5030439819 @default.
- W2933596316 date "1988-01-01" @default.
- W2933596316 modified "2023-09-23" @default.
- W2933596316 title "An accurate, efficient, and flexible algorithm for the trichotomous probit model" @default.
- W2933596316 hasPublicationYear "1988" @default.
- W2933596316 type Work @default.
- W2933596316 sameAs 2933596316 @default.
- W2933596316 citedByCount "0" @default.
- W2933596316 crossrefType "journal-article" @default.
- W2933596316 hasAuthorship W2933596316A5014766715 @default.
- W2933596316 hasAuthorship W2933596316A5030439819 @default.
- W2933596316 hasConcept C105795698 @default.
- W2933596316 hasConcept C11413529 @default.
- W2933596316 hasConcept C120784921 @default.
- W2933596316 hasConcept C140331021 @default.
- W2933596316 hasConcept C149782125 @default.
- W2933596316 hasConcept C184314375 @default.
- W2933596316 hasConcept C33923547 @default.
- W2933596316 hasConcept C41008148 @default.
- W2933596316 hasConcept C46704056 @default.
- W2933596316 hasConcept C67257552 @default.
- W2933596316 hasConceptScore W2933596316C105795698 @default.
- W2933596316 hasConceptScore W2933596316C11413529 @default.
- W2933596316 hasConceptScore W2933596316C120784921 @default.
- W2933596316 hasConceptScore W2933596316C140331021 @default.
- W2933596316 hasConceptScore W2933596316C149782125 @default.
- W2933596316 hasConceptScore W2933596316C184314375 @default.
- W2933596316 hasConceptScore W2933596316C33923547 @default.
- W2933596316 hasConceptScore W2933596316C41008148 @default.
- W2933596316 hasConceptScore W2933596316C46704056 @default.
- W2933596316 hasConceptScore W2933596316C67257552 @default.
- W2933596316 hasLocation W29335963161 @default.
- W2933596316 hasOpenAccess W2933596316 @default.
- W2933596316 hasPrimaryLocation W29335963161 @default.
- W2933596316 hasRelatedWork W1505658563 @default.
- W2933596316 hasRelatedWork W1579250242 @default.
- W2933596316 hasRelatedWork W1984684068 @default.
- W2933596316 hasRelatedWork W2005983648 @default.
- W2933596316 hasRelatedWork W2077024502 @default.
- W2933596316 hasRelatedWork W2116480496 @default.
- W2933596316 hasRelatedWork W2767528717 @default.
- W2933596316 hasRelatedWork W2895063403 @default.
- W2933596316 hasRelatedWork W2903108200 @default.
- W2933596316 hasRelatedWork W2940074097 @default.
- W2933596316 hasRelatedWork W2998820013 @default.
- W2933596316 hasRelatedWork W3082454357 @default.
- W2933596316 hasRelatedWork W3121884553 @default.
- W2933596316 hasRelatedWork W3122764916 @default.
- W2933596316 hasRelatedWork W3124897287 @default.
- W2933596316 hasRelatedWork W3137897929 @default.
- W2933596316 hasRelatedWork W3149316218 @default.
- W2933596316 hasRelatedWork W580725527 @default.
- W2933596316 hasRelatedWork W59596267 @default.
- W2933596316 hasRelatedWork W628882070 @default.
- W2933596316 isParatext "false" @default.
- W2933596316 isRetracted "false" @default.
- W2933596316 magId "2933596316" @default.
- W2933596316 workType "article" @default.