Scribd
Upload
73 views

S Weave Example

This document demonstrates how to use the ergmTable and relogitTable functions to generate LaTeX tables from regression objects estimated by the Zelig and statnet packages. It shows an example of using relogitTable to output a table from a rare-event logit model, and examples of using ergmTable to output tables from exponential random graph models estimated on a simulated network. The ergmTable function can also output tables for multiple models estimated with increasing complexity.

Uploaded by

gaby-01
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
73 views

S Weave Example

This document demonstrates how to use the ergmTable and relogitTable functions to generate LaTeX tables from regression objects estimated by the Zelig and statnet packages. It shows an example of using relogitTable to output a table from a rare-event logit model, and examples of using ergmTable to output tables from exponential random graph models estimated on a simulated network. The ergmTable function can also output tables for multiple models estimated with increasing complexity.

Uploaded by

gaby-01
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Demonstration of the ergmTable and relogitTable Functions

Thomas Jensen ETH Zrich u thomas.jensen@eup.gess.ethz.ch


First read in the libraries to be used: > library(MASS) > library(Zelig) > library(statnet) Then source the two r scripts. If you want to compile this yourself, remember to place the scripts in the same directory as the .Rnw le. > source("relogitTable.r") > source("ergmTable.r")

Rare-Event Logit Models

This section demonstrate how it works for rare-event logit models estimated by the function from the Zellig package. First we generate some fake data and dichotomize one of the variables: > > > > covvar <- matrix(c(1, 0.6, 0.6, 0.6, 1, 0.6, 0.6, 0.6, 1), nr = 3) means <- c(20, 25, 30) data <- as.data.frame(mvrnorm(100, means, covvar)) data$V1 <- ifelse(data$V1 > 20, 1, 0)

This data is not rare-event data, but is simply used to show how the script will generate a nice latex table from a zelig regression object. Next we run the rare-event logit model: > model <- zelig(V1 ~ V2 + V3, data = data, model = "relogit")
A Then the relogitTable function is used to create a L TEXtable from the above object:

> relogitTable(model)

\begin{tabular}{lD{.}{.}{3}} \toprule & \multicolumn{ 1 }{ c }{ Model 1 } \\ \midrule (Intercept) & -51.068 ^* \\ & ( 11.763 ) \\ V2 & 0.957 ^* \\ & ( 0.425 ) \\ V3 & 0.910 ^* \\ & ( 0.339 ) \\ \midrule $N$ & \multicolumn{1}{c}{100 }\\ $AIC$ & \multicolumn{1}{c}{103.815} \bottomrule \end{tabular} Which produces the following table: Model 1 (Intercept) V2 V3 N AIC 51.068 (11.763) 0.957 (0.425) 0.910 (0.339) 100 103.815

Exponential Random Graph Models

This section demonstrate how to use the the ergmTable function with ergm objects from the statnet package. First we generate a random network with twenty nodes: > net <- rgraph(20) Then we estimate a simple exponential random graph model with terms for edges, out- and in-stars and triangles: > erg <- ergm(net ~ edges + istar(1:2) + ostar(1:2) + triangle, + eval.loglik = TRUE) The ergmTable function generates the following output from the above object: > ergmTable(erg)

\begin{tabular}{lD{.}{.}{2}} \toprule & \multicolumn{ 1 }{ c }{ Model 1 } \\ \midrule edges & 1.52 \\ & ( 2.22 )\\ istar1 & -0.21 \\ & ( 1.24 )\\ istar2 & -0.09 \\ & ( 0.10 )\\ ostar1 & -0.14 \\ & ( 1.24 )\\ ostar2 & -0.21 \\ & ( 0.13 )\\ triangle & 0.09 \\ & ( 0.05 ) \\ \midrule $AIC$ & \multicolumn{1}{c}{534.05}\\ $BIC$ & \multicolumn{1}{c}{557.69} \bottomrule \end{tabular} Which produces the following table: Model 1 edges istar1 istar2 ostar1 ostar2 triangle AIC BIC 1.52 (2.22) 0.21 (1.24) 0.09 (0.10) 0.14 (1.24) 0.21 (0.13) 0.09 (0.05) 534.05 557.69

2.1

Multiple Models

Both functions also work with multiple models, here it will be demonstrated for exponential random graph models. First we estimate three models, that increase in complexity: > erg1 <- ergm(net ~ edges, eval.loglik = TRUE) > erg2 <- ergm(net ~ edges + istar(1:2) + ostar(1:2), eval.loglik = TRUE)

> erg3 <- ergm(net ~ edges + istar(1:2) + ostar(1:2) + triangle, + eval.loglik = TRUE) The ergmTable function produces the following output: > ergmTable(erg1, erg2, erg3)

\begin{tabular}{lD{.}{.}{2}@{\hspace{2em}}D{.}{.}{2}@{\hspace{2em}}D{.}{.}{2}} \toprule & \multicolumn{ 1 }{ c }{ Model 1 } & \multicolumn{ 1 }{ c }{ Model 2 } & \multicolumn{ 1 edges & 0.04 & 1.42 ^* & 1.83 ^* \\ & ( 0.10 ) & ( 0.34 ) & ( 0.38 )\\ istar1 & & -1.76 ^* & -0.96 ^*\\ & & ( 0.34 ) & ( 0.38 )\\ istar2 & & 0.02 & -0.06 \\ & & ( 0.06 ) & ( 0.08 )\\ ostar1 & & 1.06 ^* & 0.28 \\ & & ( 0.34 ) & ( 0.38 )\\ ostar2 & & -0.09 & -0.25 \\ & & ( 0.09 ) & ( 0.14 )\\ triangle & & & 0.09 \\ & & & ( 0.05 ) \\ \midrule $AIC$ & \multicolumn{1}{c}{528.62} & \multicolumn{1}{c}{534.59} & \multicolumn{1}{c}{534.40 $BIC$ & \multicolumn{1}{c}{532.56} & \multicolumn{1}{c}{554.29} & \multicolumn{1}{c}{558.04} \end{tabular} Which produces the following table: Model 1 edges istar1 istar2 ostar1 ostar2 triangle AIC BIC 528.62 532.56 534.59 554.29 0.04 (0.10) Model 2 1.42 (0.34) 1.76 (0.34) 0.02 (0.06) 1.06 (0.34) 0.09 (0.09) Model 3 1.83 (0.38) 0.96 (0.38) 0.06 (0.08) 0.28 (0.38) 0.25 (0.14) 0.09 (0.05) 534.40 558.04

You might also like