rem (version 1.1.2)

remRate: Rate function for relational event models

Description

Estimate the rate function for relational event models.

Usage

remRate(formula, 
    dist = 'weibull', 
    var.names = NULL,  
    event.sequence.type = 'continuous', ...)

Arguments

formula

A formula expression in the conventional R syntax used for linear modelling. The response (i.e. the dependent variable) must be the event sequence. On the right-hand side, covariates are listed with a +. It is possible to use one of the endogenous statistics functions in the rem-package as independent (or control) variable. flexsurvreg is called to estimate the prameters of the model for now.

dist

The distribution specified for the parametric model. See flexsurvreg for more details.

var.names

An optional string vector containing variable names for the proposed independent variables in the formula. To be used if endogenous network effects are calculated within the remRate-function.

event.sequence.type

'continuous' or 'ordinal'. Only 'continuous' is implemented as of now.

...

Optional arguments for flexsurvreg.

Details

The function remRate() allows users to estimate the rate function of the relational event model. After the variables are prepared, a flexsurvreg is called to estimate the parameters of the model.

See Also

rem-package

Examples

# create some data two-mode network event sequence data with
# a 'sender', 'target' and a 'time'-variable
sender <- c('A', 'B', 'A', 'C', 'A', 'D', 'F', 'G', 'A', 'B',
            'B', 'C', 'D', 'E', 'F', 'B', 'C', 'D', 'E', 'C', 
            'A', 'F', 'E', 'B', 'C', 'E', 'D', 'G', 'A', 'G', 
            'F', 'B', 'C')
target <- c('T1', 'T2', 'T3', 'T2', 'T1', 'T4', 'T6', 'T2', 
            'T4', 'T5', 'T5', 'T5', 'T1', 'T6', 'T7', 'T2', 
            'T3', 'T1', 'T1', 'T4', 'T5', 'T6', 'T8', 'T2',
            'T7', 'T1', 'T6', 'T7', 'T3', 'T4', 'T7', 'T8', 'T2')
time <- c('03.01.15', '04.01.15', '10.02.15', '28.02.15', '01.03.15', 
          '07.03.15', '07.03.15', '12.03.15', '04.04.15', '28.04.15',
          '06.05.15', '11.05.15', '13.05.15', '17.05.15', '22.05.15', 
          '09.08.15', '09.08.15', '14.08.15', '16.08.15', '29.08.15',
          '05.09.15', '25.09.15', '02.10.15', '03.10.15', '11.10.15', 
          '18.10.15', '20.10.15', '28.10.15', '04.11.15', '09.11.15', 
          '10.12.15', '11.12.15', '12.12.15')
type <- sample(c('con', 'pro'), 33, replace = TRUE)
important <- sample(c('important', 'not important'), 33,
                    replace = TRUE)
sender.country <- sample(c('country1', 'country2'), 33, 
                         replace = TRUE)

# combine them into a data.frame
dt <- data.frame(sender, target, time, type, important, sender.country)

# create event sequence and order the data
dt <- eventSequence(datevar = dt$time, dateformat = '%d.%m.%y', 
                    data = dt, type = 'continuous', 
                    byTime = 'daily', returnData = TRUE,
                    sortData = TRUE)

# create some endogenous variables: 
dt$inertia <- inertiaStat(dt, dt$event.seq.cont, dt$sender, dt$target, 30)
dt$activitySender <- degreeStat(dt, dt$event.seq.cont, dt$sender, 30)
dt$targetPopularity <- degreeStat(dt, dt$event.seq.cont,dt$target, 30)

# check out their correlation coefficient
cor(subset(dt, select = c(inertia, activitySender, 
                          targetPopularity)), method="pearson") 

# fit event history model using inertia, actor activity and
# target popularity 
fit1 <- remRate(
  dt$event.seq.cont ~ 
    inertiaStat(dt, dt$event.seq.cont, dt$sender, dt$target, 30) +
    degreeStat(dt, dt$event.seq.cont, dt$sender, 30) +
    degreeStat(dt, dt$event.seq.cont,dt$target, 30),
  var.names = c('inertia', 'actorAct', 'targetPop'))
fit1
## Interpretation of the weibull-model coefficients:
# For positive coefficients, if the independent variable increases 
# in value, the survival time increases, meaning the event will occurr 
# at a lower rate. The coefficients are log odds. Calculate the odds 
# ratios by runnding exp(coefficient). This will give you the factor 
# by which the risk of the event occurring increases 
# (negative coefficient, OR < 1) or decreases 
# (positive coefficient, OR > 1).

# calculate closing four-cycles
dt$fourCyc <- fourCycleStat(dt, dt$event.seq.cont, 
                            dt$sender, dt$target, 30)
dt$fourCyc.support <- fourCycleStat(dt, dt$event.seq.cont, 
                                    dt$sender, dt$target, 30, 
                                    eventtypevar = dt$type, 
                                    eventtypevalue = 'positive')

# do actors from country1 engage more often in positive
# closing four-cycles?
library('ggplot2')
ggplot(dt, aes (event.seq.cont, fourCyc.support, 
                color = sender.country)) +
  geom_point()+ geom_smooth() 

# test above idea of closing four-cycles with an 
# interaction term with 'sender.country'
fit2 <- remRate(
  dt$event.seq.cont ~ 
    inertiaStat(dt, dt$event.seq.cont, dt$sender, dt$target, 30) +
    degreeStat(dt, dt$event.seq.cont, dt$sender, 30) +
    degreeStat(dt, dt$event.seq.cont,dt$target, 30) +
    dt$fourCyc.support*dt$sender.country,
  var.names = c('inertia', 'actorActivity', 
                'targetPopularity', 
                'fourCyc.support', 'country',
                'fourCyc.supportXcountry'))
fit2

# sender similarity: how likely is it that two
# senders are alike?
fit3 <- remRate(
  dt$event.seq.cont ~ 
    inertiaStat(dt, dt$event.seq.cont, dt$sender, dt$target, 30) +
    degreeStat(dt, dt$event.seq.cont, dt$sender, 30) +
    degreeStat(dt, dt$event.seq.cont,dt$target, 30) +
    similarityStat(dt, dt$event.seq.cont, dt$sender, dt$target,
                   halflife.last.event = 10),
  var.names = c('inertia', 'actorActivity', 
                'targetPopularity', 'simSender'))
fit3

# target similarity: how likely is it that two
# targets are used together (e.g. the current actor 
# copies others/learns from others)
fit4 <- remRate(
  dt$event.seq.cont ~ 
    inertiaStat(dt, dt$event.seq.cont, dt$sender, dt$target, 30) +
    degreeStat(dt, dt$event.seq.cont, dt$sender, 30) +
    degreeStat(dt, dt$event.seq.cont,dt$target, 30) +
    similarityStat(dt, dt$event.seq.cont, dt$sender, dt$target,
                   senderOrTarget = 'target',
                   halflife.last.event = 10),
  var.names = c('inertia', 'actorActivity', 
                'targetPopularity', 'simTarget'))
fit4