Notice: On March 31, it was **announced** that Statalist is moving from an email list to a **forum**. The old list will shut down at the end of May, and its replacement, **statalist.org** is already up and running.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

From |
hdneder@ufu.br |

To |
statalist@hsphsun2.harvard.edu |

Subject |
st: predictions out of sample with spatial regression |

Date |
Fri, 21 Jun 2013 00:16:24 -0300 (BRT) |

Dear Stata List members I have a sample of n observations of the dependent variable y and independent variables X1, X2, ..., Xk. I wish to estimate a spatial lag model for a sample of n observations with the following specification: y = rho * W * X * beta + y + u where W is a neighborhood binary matrix (or inverse distances matrix) and rho is a spatial autoregressive parameter. I realize the estimation of this model for n observations and wish to make predictions within and outside the sample. To make estimates within the sample is simple, but how to realize the estimation of a set of observations (out of sample) that surround the original n observations? One idea is to rebuild the neighborhood matrix for each point outside the sample and from each new matrix W '(with one more row and column than W) I would apply the expression. with the same parameters estimates: ypred = rho*W'y' + beta *X Assuming this space is structural homogeneous, this procedure is valid? If it is true, are there any Stata command to automatically perform these predictions out of sample or I have to develop a specific routine for this? But I guess that have some problems with this procedure. Any help is wellcome. Best regards Henrique Neder Google Tradutor para empresas:Google Translator ToolkitTradutor de sitesGlobal Market Finder Desativar tradução instantâneaSobre o Google TradutorCelular * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/faqs/resources/statalist-faq/ * http://www.ats.ucla.edu/stat/stata/

- Prev by Date:
**Re: st: Discrete time hazard models using cloglog and svy** - Next by Date:
**Re: st: Missing observations** - Previous by thread:
**st: Sample size for four-level logistic regression** - Next by thread:
**st: error codes returned by NL function evaluator program to use dummy variables** - Index(es):