search for: limdep

Hi all, Sorry to bother you. I'm estimating a discrete choice model in R using the maxBFGS command. Since I wrote the log-likelihood myself, in order to double check, I run the same model in Limdep. It turns out that the coefficient estimates are quite close; however, the standard errors are very different. I also computed the hessian and outer product of the gradients in R using the numDeriv package, but the results are still very different from those in Limdep. Is it the routine to compute...

Dear R users, I would like to perform Latent Class Logit Models for the analysis of choice experiments in environmental valuation. This kind of analysis is usually performed with NLogit Software (http://www.limdep.com). I attach the results I usually obtain using NLogit and NLogit model specifications. For Random parameter models and Logit Models I usually perform my analysis with the package mlogit ( http://cran.r-project.org/web/packages/mlogit/index.html ). The models I would like to run are presented...

...s. This is for teaching and expository purposes only. I've successfully programmed the normal, generalized gamma, gamma, weibull, exponential, and lognormal regression functions. And optim returns reasonable answers for all of these compared with the identical optimization problems in STATA and LIMDEP. However, I'm having a problem with beta regression. The parameterization of beta regression that I am using solves for the second distributional parameter in terms of the first and xbeta. Therefore, there is only a single distribution parameter. This parameterization works fine in STATA, LIMDE...

...survey dataset that contains 2,000 individuals and one of the survey questions asked the respondents to chose two main reasons to work out of eight options (e.g., pay, pepople, to use abilities, etc.) . Now I'd like to model the responses using simultaneous logit/probit models. To my knowledge, LIMDEP (authored by Greene) can handle these models. As I know little about LIMDEP, I wonder if R has any package designed for these models. I have searched the R archive but haven't got any clue. Any suggestions? Thanks, Min ------------------------- Sociology Dept. Oxford University

...multino- > mial logit is used here to denote the model used by Schmidt and > Strauss (1975), while > multinomial logit is used as a synonym of conditional logit. I found this web site useful in comparing various categorical dependent variable models, but it only addresses SAS, STATA, LIMDEP, and SPSS: http://www.indiana.edu/~statmath/stat/all/cdvm/cdvm1.html I have searched using: 1. Google> site:r-project multinomial nested logit 2. Google> site:r-project.org multinomial discrete choice 3. R> help.search("multinomial nested logit") No help files found matching ?...

Dear all: I'm trying to fit the optimal Box-Cox transformation related to nls (see the code below) for the demand of money data in Green (3th Edition) but in the last step R gives the next error message. Error en `[.data.frame`(eval(object$data), , as.character(formula(object)[[2]])[2]) : undefined columns selected. ?Any idea to solve the problem? Thanks in advance,

Hi -- subject says all. I just want the code for that function, which I guess was in Venables and Ripley as early as 1994. Well, and for any of the sub-functions that glm.nb calls. I can't install the entire MASS library. If the code for just glm.nb (again, don't want to touch the MASS library, last time I tried to install it was a complete nightmare and fiasco) is somewhere on a

...hat is the dependent variable >0 and upper truncated at 1; bhat0 is the intercept; Z is the independent variable and is a uniform random variate between 0 and 1; and e is the error term. I realised that R doesn't have a built-in function to estimate truncated regression models as does STATA, LIMDEP etc. I tried the survival and FEAR packages and couldn't fit it for my case. So I wrote the log likelihood function of the truncated regression model and reparametrised it using Olsen (1978) so that the function is globally concave and has an unique maximiser. I used a quasi-Newton method (BFGS...

...s where one coefficient is random and normally distributed. I've searched on the R help archive and don't see much information very specific to what I'm doing, so I write the code myself, which involves simulated maximum likelihood. But it doesn't work, as I compare the results with Limdep. Could anyone please take a look at it? It's a bit longer though......Comments are highly appreciated. Thank you very much! /*1,000 people, Three alternatives, two variables, their coefficients are B1 and B2 respectively, B1 is random*/ library(maxLik) RP<-function(theta,y,X) { m...

A colleague insalled Wine for me on a 32-bit Red Hat Linux server yesterday. I am not trying to install an application called LIMDEP under Wine. The documentation says to start a script called winecfg. When I invoke that script, all I get is a gray box on my screen with a few tabs. No text appears on the screen. If I click on some of the tabs, there are drop-down menus, which also have no text. None of the buttons have any text...

Dear R users, I am happy to announce the initial release of the "mvProbit" package on CRAN (version 0.1-0). This package provides tools for econometric analysis with Multivariate Probit Models. While these models can be estimated also by several other statistical software packages (e.g. LIMDEP/NLOGIT, STATA), "mvProbit" is much more flexible and powerful in calculating marginal effects. To my best knowledge, "mvProbit" is the only statistical software package that can calculate various marginal effects including their standard errors at arbitrary user-defined values o...

Dear R users, I am happy to announce the initial release of the "mvProbit" package on CRAN (version 0.1-0). This package provides tools for econometric analysis with Multivariate Probit Models. While these models can be estimated also by several other statistical software packages (e.g. LIMDEP/NLOGIT, STATA), "mvProbit" is much more flexible and powerful in calculating marginal effects. To my best knowledge, "mvProbit" is the only statistical software package that can calculate various marginal effects including their standard errors at arbitrary user-defined values o...

...erogeneous negative binomial. I have written a function called "nbinomial" which is in the msme package on CRAN. Type ?nbinomial to see the help file. The default model is a negative binomial for which the dispersion parameter is directly related to mu, which is how Stata, SAS, SPSS, Limdep, and so forth parameterize the negative binomial. The direct parameterization make sense in that the more variation or correlation there is in a Poisson model, the greater is the value of the dispersion parameter which is adjusting for the excessive variation. With this parameterization the d...