networkDynamic (version 0.11.4)

net.obs.period: Network attribute describing the observation properties of a networkDynamic object

Description

net.obs.period is an optional network-level attribute that gives meta-information about the characteristics of the observations, time range and temporal model of a networkDynamic object.

Format

If the attribute is present, it is required to have a specific structure. It is a 4-element list with named values.

• the observations element is a list of two-elment spell vectors that give the onset and terminus of the time period(s) during which the network was observed or simulated

• mode A string indicating the temporal model of the network: either discrete or continuous

• time.increment is an indicator of the `natural' time unit for iterating over the network (primarily useful for discrete time models)

• time.unit is the name of the unit used by all of the spells in the object: often set to step for discrete simulations.

Details

Some functions that work with networkDynamic objects need to know information about the observation design of the network. Is it discrete or continuous time model? When did the observation (or simulation) of the network processes start and end? What are the time units? This information may be stored in the optional net.obs.period network attribute.

Some examples of possible net.obs.period values for various networks:

• for an stergm output it might be: list(observations=list(c(0,100)),mode="discrete", time.increment=1,time.unit="step")

• For Lin's windsurfers , which is missing all observations on day 25: list(observations=list(c(0,25),c(26,31)),mode="discrete", time.increment=1,time.unit="day")

• For McFarland's classroom data: list(observations=list(c(0,55),),mode="continuous", time.increment=NA,time.unit="minute")

• A fictitious instantaneous panel example : list(observations=list(c(0,0),c(21,21),c(56.5,56.5),c(68,68)),mode="discrete", time.increment=0,time.unit="seconds")

  The values of the net.obs.period object are not checked for consistency with activity spells of the network itself.

References

Link to page on statnet wiki?

See Also

Several of the networkDynamic converter functions set the net.obs.period attribute on their output object. See also activity.attribute

Examples

nd<-network.initialize(5)
activate.vertices(nd,onset=0,terminus=100)
set.network.attribute(nd,'net.obs.period',list(observations=list(c(0,100)),
                      mode="discrete", time.increment=1,time.unit="step"))
# find the overall `expected' time range for the object
range(get.network.attribute(nd,'net.obs.period')$observations)