[R-sig-Geo] Model - SAR Poisson GLM (Ticiana Grecco Zanon)

[Ticiana Grecco Zanon]

Hello Corey,

Yes, this is what I needed! Thank you so much, I will prove it with my data.

Regards,
Ticiana Grecco Zanon Moura
UESC - Universidade Estadual de Santa Cruz
http://www.uesc.br
REGIOlab - Laboratorio de Análisis Económico Regional
http://www.uniovi.net/regiolab
________________________________


De: Corey Sparks <corey.sparks at utsa.edu>
Enviado: terça-feira, 7 de março de 2017 12:46:19
Para: r-sig-geo at r-project.org
Cc: ticianaz at hotmail.com
Assunto: RE: Model - SAR Poisson GLM (Ticiana Grecco Zanon)

Hello, there is such a thing as an auto-Poisson model. I have an rpub here
http://rpubs.com/corey_sparks/250410

and towards the bottom, I show how to lag the count outcome and include it in the model. You could do this with any variable in your model (create a lag of it), by using the lag.listw() function in spdep.

You can also include the lag term in a negative binomial if you want to include overdispersion, without resorting to the 'quasi' distributions.
Best,
Corey

Corey Sparks, Ph.D.
Associate Professor
Graduate Advisor of Record
Department of Demography
The University of Texas at San Antonio
501 West Cesar E Chavez Blvd
corey.sparks 'at' utsa.edu<http://t.sidekickopen55.com/e1t/c/5/f18dQhb0S7lC8dDMPbW2n0x6l2B9nMJW7t5XZs3LQ7cCN63RMPbd7cf-W8qlRZx56dPL4f2wwDQl02?si=5625354980425728&pi=3F5E8B84-79AC-4CDA-8B3A-E2CA0310971B&t=http%3A%2F%2Futsa%2Eedu>
coreysparks.weebly.com<http://t.sidekickopen55.com/e1t/c/5/f18dQhb0S7lC8dDMPbW2n0x6l2B9nMJW7t5XZs3LQ7cCN63RMPbd7cf-W8qlRZx56dPL4f2wwDQl02?si=5625354980425728&pi=3F5E8B84-79AC-4CDA-8B3A-E2CA0310971B&t=http%3A%2F%2Fcoreysparks%2Eweebly%2Ecom>