Lab

install.packages("statnet")   # Network hypotheses/models
install.packages("ergm")      # ERG models (in Statnet)
install.packages("Rglpk")     # Needed for model estimation

#Import matrix
data <- read.csv("~/Desktop/Sunbelt_June/Sunbelt_Ville_June/adjacency_matrix.csv", header=F, as.is = F)


# ================ ~~ Directed network ================
help('ergm-terms')

library(statnet)

#Create two Statnet network objects

#---------Convert into matrix---------#
data_stat <- as.matrix(data)

#---------Create Statnet network object---------#
#network object
data_net <- network(data_stat, directed = TRUE, matrix.type = "adjacency",
                    loops=F, multiple=F)
data_net


#---------Exponential random graph model---------#
#install.packages("ergm")
library(ergm)

help('ergm-terms')

#1 
lap_model <- ergm(data_net ~ edges)
summary(lap_model)

#2 
lap_model <- ergm(data_net ~ edges + mutual)  
summary(lap_model)

#1-5
lap_model <- ergm(data_net ~ edges #connections 
                   + mutual #reciprocity 
                   + gwesp(0.2, fixed=T) #Transitive closure 
                   + edgecov(doc_net) #H4: Doc sharing tie 
                   + nodeicov("expertise") #H5: Incoming links
                   + nodeocov("expertise") #H6: Outgoing links
)
summary(lap_model)



#---------Goodness of Fit---------#

#Calculate gof and plot 
gof_1 <- gof(lap_model)
gof_1
plot(gof_1)