Risk set
This vignette provides a definition of the full,
active, active_saturated, and manual risk
sets, explains how each is specified in remify::remify(),
and shows how the processed risk set is represented in the returned
remify object.
Consider a small example network:
library(remify) # loading package
data(randomREHsmall) # data
# processing the edgelist
reh <- remify(
edgelist = randomREHsmall$edgelist,
directed = TRUE,
ordinal = FALSE,
model = "tie",
actors = randomREHsmall$actors,
origin = randomREHsmall$origin
)Definition of risk set
A relational event history consists of a time-ordered sequence of (directed or undirected) interactions. For each event, we know:
- its time of occurrence, either as a timestamp/date/continuous value or as an order index
- the actors involved
- the type of the event (if measured)
For instance, the first five events of the
randomREHsmall sequence are:
## time actor1 actor2
## 1 2020-03-05 16:36:37 Colton Kayla
## 2 2020-03-05 19:34:11 Lexy Colton
## 3 2020-03-05 20:49:37 Colton Kayla
## 4 2020-03-05 21:38:23 Colton Kayla
## 5 2020-03-06 06:54:12 Richard Coltonwhere time, actor1, actor2
describe each observed event (in this example, event types are not
annotated).
When modeling a relational event sequence, we define per each time point a risk set: the set of dyadic events that were at risk of occurring at that time point (which always includes the event that actually occurred). The risk set is a central building block of the likelihood for both tie-oriented and actor-oriented models. Four possible definitions are discussed below: full, active, active_saturated, and manual.
The full risk set
The full risk set assumes all possible dyads are at risk throughout the entire observation period. For a network with \(N\) actors, \(C\) event types, and directed events, the risk set has size \(D = N(N-1)C\); for undirected events, \(D = N(N-1)C/2\).
In randomREHsmall, dyads are directed, \(N = 5\) actors, and \(C = 1\) (untyped), so \(D = 5 \times 4 \times 1 = 20\).
The risk set dyad composition is stored in
reh$index$dyad_map:
## dyadID actor1 actor2
## 1 1 Colton Francesca
## 2 2 Colton Kayla
## 3 3 Colton Lexy
## 4 4 Colton Richard
## 5 5 Francesca Colton
## 6 6 Francesca KaylaThe ID of each dyad is defined by a two-step approach:
Step 1. Actors’ names are sorted alphanumerically and assigned integer IDs from \(1\) to \(N\):
## [1] "Colton" "Francesca" "Kayla" "Lexy" "Richard"## 1 2 3 4 5
## "Colton" "Francesca" "Kayla" "Lexy" "Richard"Step 2. Dyads are indexed by looping first over
actor1, then actor2, and finally
type (if applicable), skipping self-loops:
sorted_types <- c(" ") # no event type in randomREHsmall
C <- length(sorted_types)
dyad_mat <- matrix(NA, nrow = N*(N-1)*C, ncol = 3)
colnames(dyad_mat) <- c("actor1", "actor2", "type")
rownames(dyad_mat) <- 1:(N*(N-1)*C)
d <- 1
for(type in sorted_types){
for(actor1 in sorted_actors){
for(actor2 in sorted_actors){
if(actor1 != actor2){
dyad_mat[d,] <- c(actor1, actor2, type)
d <- d + 1
}
}
}
}
dyad_mat[1:5,]## actor1 actor2 type
## 1 "Colton" "Francesca" " "
## 2 "Colton" "Kayla" " "
## 3 "Colton" "Lexy" " "
## 4 "Colton" "Richard" " "
## 5 "Francesca" "Colton" " "## [1] 20The row index in dyad_mat corresponds to the
dyadID. This ID is used throughout ‘remverse’ to identify
dyads. For tie-oriented modeling, the observed dyad IDs per event are
stored in reh$ids$dyad:
## [1] 2 13 2 2 17 2Visualizing the risk set
A useful visualization plots the risk set as a grid with senders on the y-axis and receivers on the x-axis. At each time point, the observed dyad is colored green and all other at-risk dyads are colored gray.
Considering the first four time points of
randomREHsmall, we observe: the dyad (Colton,
Kayla) at \(t_1\), \(t_3\), and \(t_4\), and the dyad (Lexy, Colton)
at \(t_2\). Green cells mark observed
events; gray cells mark other at-risk dyads; white cells mark infeasible
events (self-loops).
For undirected networks, dyads are sorted alphanumerically before
processing, so the risk set is restricted to the lower triangular grid.
For instance, c("Lexy", "Colton") becomes
c("Colton", "Lexy").
The active, active_saturated, and manual risk set
The full risk set maintains a constant structure throughout the event history: all possible dyads are always at risk. Three alternative definitions accommodate settings where not all dyads are plausibly at risk.
The active risk set
In sparse networks, the number of observed dyads can be far smaller
than \(D\). The active risk
set restricts the risk set to only those dyads observed at least once in
the event history, and maintains this structure throughout. Specify with
riskset = "active".
The use of the active risk set can substantially decrease computational time for both statistics and model estimation. However, this reduction excludes dyads that were never observed but might be plausibly at risk. It is good practice to examine the active dyads before proceeding and to consider whether the exclusion of unobserved dyads is defensible for the data at hand.
reh_active <- remify(
edgelist = randomREHsmall$edgelist,
directed = TRUE,
ordinal = FALSE,
model = "tie",
actors = randomREHsmall$actors,
riskset = "active",
origin = randomREHsmall$origin
)
# number of dyads in full vs active risk set
reh_active$D## [1] 20## [1] 20## actor1 actor2 dyadID actor1ID actor2ID
## 1 Colton Francesca 1 1 2
## 2 Colton Kayla 2 1 3
## 3 Colton Lexy 3 1 4
## 4 Colton Richard 4 1 5
## 5 Francesca Colton 5 2 1
## 6 Francesca Kayla 6 2 3
## 7 Francesca Lexy 7 2 4
## 8 Francesca Richard 8 2 5
## 9 Kayla Colton 9 3 1
## 10 Kayla Francesca 10 3 2
## 11 Kayla Lexy 11 3 4
## 12 Kayla Richard 12 3 5
## 13 Lexy Colton 13 4 1
## 14 Lexy Francesca 14 4 2
## 15 Lexy Kayla 15 4 3
## 16 Lexy Richard 16 4 5
## 17 Richard Colton 17 5 1
## 18 Richard Francesca 18 5 2
## 19 Richard Kayla 19 5 3
## 20 Richard Lexy 20 5 4The grid visualization below shows the active risk set at time \(t_3\). Only the dyads observed at least once in the history (gray and green tiles) are at risk; all other dyads are excluded (white tiles).
The active_saturated risk set
The active_saturated risk set extends the active risk set by
additionally including the reverse direction of each observed dyad and
all event types for each observed actor pair. For instance, if A→B is
observed, B→A is also included in the risk set. This reflects the
assumption that observing an interaction between two actors implies that
both directions (and all event types) are structurally possible, even if
not yet observed. Specify with
riskset = "active_saturated".
reh_sat <- remify(
edgelist = randomREHsmall$edgelist,
directed = TRUE,
model = "tie",
riskset = "active_saturated",
actors = randomREHsmall$actors,
origin = randomREHsmall$origin
)
reh_sat$activeD## [1] 20## actor1 actor2 dyadID actor1ID actor2ID
## 1 Colton Francesca 1 1 2
## 2 Colton Kayla 2 1 3
## 3 Colton Lexy 3 1 4
## 4 Colton Richard 4 1 5
## 5 Francesca Colton 5 2 1
## 6 Francesca Kayla 6 2 3
## 7 Francesca Lexy 7 2 4
## 8 Francesca Richard 8 2 5
## 9 Kayla Colton 9 3 1
## 10 Kayla Francesca 10 3 2
## 11 Kayla Lexy 11 3 4
## 12 Kayla Richard 12 3 5
## 13 Lexy Colton 13 4 1
## 14 Lexy Francesca 14 4 2
## 15 Lexy Kayla 15 4 3
## 16 Lexy Richard 16 4 5
## 17 Richard Colton 17 5 1
## 18 Richard Francesca 18 5 2
## 19 Richard Kayla 19 5 3
## 20 Richard Lexy 20 5 4The manual risk set
There are settings where only a specific, user-defined set of dyads
should be considered at risk throughout the event history. This is the
manual risk set, specified via riskset = "manual"
and the manual.riskset argument.
manual.riskset is a data.frame with columns
actor1 and actor2 (and optionally
type) listing all dyads that are at risk. This defines a
static risk set that does not vary over time. Any
observed dyads from the edgelist that are absent from
manual.riskset are added automatically with a warning.
Typical use cases include:
- Structural constraints: only certain actor pairs can interact (e.g., members of specific organizational units, or actors within each other’s contact list).
- Sparse networks: the user has domain knowledge about which dyads are plausibly at risk, beyond just those observed.
- Spatial or group constraints: only actors present in the same location or group can interact.
Specifying a manual risk set
data(randomREH)
# Suppose we want to restrict the risk set to dyads among three actors only,
# while all other dyads are considered structurally impossible.
manual_rs <- data.frame(
actor1 = c("Alexander", "Alexander", "Colton", "Colton", "Lexy", "Lexy"),
actor2 = c("Colton", "Lexy", "Alexander", "Lexy", "Alexander", "Colton")
)
reh_manual <- remify(
edgelist = randomREH$edgelist,
directed = TRUE,
ordinal = FALSE,
model = "tie",
actors = randomREH$actors,
riskset = "manual",
manual.riskset = manual_rs,
origin = randomREH$origin
)## Note: `actors` is ignored when `riskset = "manual"` and `manual.riskset` is used## Warning in remify(edgelist = randomREH$edgelist, directed = TRUE, ordinal =
## FALSE, : 1112 observed dyad-type combinations were added to the manual risk
## set.## [1] 380## actor1 actor2 dyadID actor1ID actor2ID
## 1 Alexander Andrey 1 1 2
## 2 Alexander Breanna 2 1 3
## 3 Alexander Charles 3 1 4
## 4 Alexander Colton 4 1 5
## 5 Alexander Crystal 5 1 6
## 6 Alexander Derek 6 1 7
## 7 Alexander Francesca 7 1 8
## 8 Alexander Justin 8 1 9
## 9 Alexander Kayla 9 1 10
## 10 Alexander Kelsey 10 1 11
## 11 Alexander Kiffani 11 1 12
## 12 Alexander Lexy 12 1 13
## 13 Alexander Maya 13 1 14
## 14 Alexander Mckenna 14 1 15
## 15 Alexander Megan 15 1 16
## 16 Alexander Michaela 16 1 17
## 17 Alexander Richard 17 1 18
## 18 Alexander Wyatt 18 1 19
## 19 Alexander Zackary 19 1 20
## 20 Andrey Alexander 20 2 1
## 21 Andrey Breanna 21 2 3
## 22 Andrey Charles 22 2 4
## 23 Andrey Colton 23 2 5
## 24 Andrey Crystal 24 2 6
## 25 Andrey Derek 25 2 7
## 26 Andrey Francesca 26 2 8
## 27 Andrey Justin 27 2 9
## 28 Andrey Kayla 28 2 10
## 29 Andrey Kelsey 29 2 11
## 30 Andrey Kiffani 30 2 12
## 31 Andrey Lexy 31 2 13
## 32 Andrey Maya 32 2 14
## 33 Andrey Mckenna 33 2 15
## 34 Andrey Megan 34 2 16
## 35 Andrey Michaela 35 2 17
## 36 Andrey Richard 36 2 18
## 37 Andrey Wyatt 37 2 19
## 38 Andrey Zackary 38 2 20
## 39 Breanna Alexander 39 3 1
## 40 Breanna Andrey 40 3 2
## 41 Breanna Charles 41 3 4
## 42 Breanna Colton 42 3 5
## 43 Breanna Crystal 43 3 6
## 44 Breanna Derek 44 3 7
## 45 Breanna Francesca 45 3 8
## 46 Breanna Justin 46 3 9
## 47 Breanna Kayla 47 3 10
## 48 Breanna Kelsey 48 3 11
## 49 Breanna Kiffani 49 3 12
## 50 Breanna Lexy 50 3 13
## 51 Breanna Maya 51 3 14
## 52 Breanna Mckenna 52 3 15
## 53 Breanna Megan 53 3 16
## 54 Breanna Michaela 54 3 17
## 55 Breanna Richard 55 3 18
## 56 Breanna Wyatt 56 3 19
## 57 Breanna Zackary 57 3 20
## 58 Charles Alexander 58 4 1
## 59 Charles Andrey 59 4 2
## 60 Charles Breanna 60 4 3
## 61 Charles Colton 61 4 5
## 62 Charles Crystal 62 4 6
## 63 Charles Derek 63 4 7
## 64 Charles Francesca 64 4 8
## 65 Charles Justin 65 4 9
## 66 Charles Kayla 66 4 10
## 67 Charles Kelsey 67 4 11
## 68 Charles Kiffani 68 4 12
## 69 Charles Lexy 69 4 13
## 70 Charles Maya 70 4 14
## 71 Charles Mckenna 71 4 15
## 72 Charles Megan 72 4 16
## 73 Charles Michaela 73 4 17
## 74 Charles Richard 74 4 18
## 75 Charles Wyatt 75 4 19
## 76 Charles Zackary 76 4 20
## 77 Colton Alexander 77 5 1
## 78 Colton Andrey 78 5 2
## 79 Colton Breanna 79 5 3
## 80 Colton Charles 80 5 4
## 81 Colton Crystal 81 5 6
## 82 Colton Derek 82 5 7
## 83 Colton Francesca 83 5 8
## 84 Colton Justin 84 5 9
## 85 Colton Kayla 85 5 10
## 86 Colton Kelsey 86 5 11
## 87 Colton Kiffani 87 5 12
## 88 Colton Lexy 88 5 13
## 89 Colton Maya 89 5 14
## 90 Colton Mckenna 90 5 15
## 91 Colton Megan 91 5 16
## 92 Colton Michaela 92 5 17
## 93 Colton Richard 93 5 18
## 94 Colton Wyatt 94 5 19
## 95 Colton Zackary 95 5 20
## 96 Crystal Alexander 96 6 1
## 97 Crystal Andrey 97 6 2
## 98 Crystal Breanna 98 6 3
## 99 Crystal Charles 99 6 4
## 100 Crystal Colton 100 6 5
## 101 Crystal Derek 101 6 7
## 102 Crystal Francesca 102 6 8
## 103 Crystal Justin 103 6 9
## 104 Crystal Kayla 104 6 10
## 105 Crystal Kelsey 105 6 11
## 106 Crystal Kiffani 106 6 12
## 107 Crystal Lexy 107 6 13
## 108 Crystal Maya 108 6 14
## 109 Crystal Mckenna 109 6 15
## 110 Crystal Megan 110 6 16
## 111 Crystal Michaela 111 6 17
## 112 Crystal Richard 112 6 18
## 113 Crystal Wyatt 113 6 19
## 114 Crystal Zackary 114 6 20
## 115 Derek Alexander 115 7 1
## 116 Derek Andrey 116 7 2
## 117 Derek Breanna 117 7 3
## 118 Derek Charles 118 7 4
## 119 Derek Colton 119 7 5
## 120 Derek Crystal 120 7 6
## 121 Derek Francesca 121 7 8
## 122 Derek Justin 122 7 9
## 123 Derek Kayla 123 7 10
## 124 Derek Kelsey 124 7 11
## 125 Derek Kiffani 125 7 12
## 126 Derek Lexy 126 7 13
## 127 Derek Maya 127 7 14
## 128 Derek Mckenna 128 7 15
## 129 Derek Megan 129 7 16
## 130 Derek Michaela 130 7 17
## 131 Derek Richard 131 7 18
## 132 Derek Wyatt 132 7 19
## 133 Derek Zackary 133 7 20
## 134 Francesca Alexander 134 8 1
## 135 Francesca Andrey 135 8 2
## 136 Francesca Breanna 136 8 3
## 137 Francesca Charles 137 8 4
## 138 Francesca Colton 138 8 5
## 139 Francesca Crystal 139 8 6
## 140 Francesca Derek 140 8 7
## 141 Francesca Justin 141 8 9
## 142 Francesca Kayla 142 8 10
## 143 Francesca Kelsey 143 8 11
## 144 Francesca Kiffani 144 8 12
## 145 Francesca Lexy 145 8 13
## 146 Francesca Maya 146 8 14
## 147 Francesca Mckenna 147 8 15
## 148 Francesca Megan 148 8 16
## 149 Francesca Michaela 149 8 17
## 150 Francesca Richard 150 8 18
## 151 Francesca Wyatt 151 8 19
## 152 Francesca Zackary 152 8 20
## 153 Justin Alexander 153 9 1
## 154 Justin Andrey 154 9 2
## 155 Justin Breanna 155 9 3
## 156 Justin Charles 156 9 4
## 157 Justin Colton 157 9 5
## 158 Justin Crystal 158 9 6
## 159 Justin Derek 159 9 7
## 160 Justin Francesca 160 9 8
## 161 Justin Kayla 161 9 10
## 162 Justin Kelsey 162 9 11
## 163 Justin Kiffani 163 9 12
## 164 Justin Lexy 164 9 13
## 165 Justin Maya 165 9 14
## 166 Justin Mckenna 166 9 15
## 167 Justin Megan 167 9 16
## 168 Justin Michaela 168 9 17
## 169 Justin Richard 169 9 18
## 170 Justin Wyatt 170 9 19
## 171 Justin Zackary 171 9 20
## 172 Kayla Alexander 172 10 1
## 173 Kayla Andrey 173 10 2
## 174 Kayla Breanna 174 10 3
## 175 Kayla Charles 175 10 4
## 176 Kayla Colton 176 10 5
## 177 Kayla Crystal 177 10 6
## 178 Kayla Derek 178 10 7
## 179 Kayla Francesca 179 10 8
## 180 Kayla Justin 180 10 9
## 181 Kayla Kelsey 181 10 11
## 182 Kayla Kiffani 182 10 12
## 183 Kayla Lexy 183 10 13
## 184 Kayla Maya 184 10 14
## 185 Kayla Mckenna 185 10 15
## 186 Kayla Megan 186 10 16
## 187 Kayla Michaela 187 10 17
## 188 Kayla Richard 188 10 18
## 189 Kayla Wyatt 189 10 19
## 190 Kayla Zackary 190 10 20
## 191 Kelsey Alexander 191 11 1
## 192 Kelsey Andrey 192 11 2
## 193 Kelsey Breanna 193 11 3
## 194 Kelsey Charles 194 11 4
## 195 Kelsey Colton 195 11 5
## 196 Kelsey Crystal 196 11 6
## 197 Kelsey Derek 197 11 7
## 198 Kelsey Francesca 198 11 8
## 199 Kelsey Justin 199 11 9
## 200 Kelsey Kayla 200 11 10
## 201 Kelsey Kiffani 201 11 12
## 202 Kelsey Lexy 202 11 13
## 203 Kelsey Maya 203 11 14
## 204 Kelsey Mckenna 204 11 15
## 205 Kelsey Megan 205 11 16
## 206 Kelsey Michaela 206 11 17
## 207 Kelsey Richard 207 11 18
## 208 Kelsey Wyatt 208 11 19
## 209 Kelsey Zackary 209 11 20
## 210 Kiffani Alexander 210 12 1
## 211 Kiffani Andrey 211 12 2
## 212 Kiffani Breanna 212 12 3
## 213 Kiffani Charles 213 12 4
## 214 Kiffani Colton 214 12 5
## 215 Kiffani Crystal 215 12 6
## 216 Kiffani Derek 216 12 7
## 217 Kiffani Francesca 217 12 8
## 218 Kiffani Justin 218 12 9
## 219 Kiffani Kayla 219 12 10
## 220 Kiffani Kelsey 220 12 11
## 221 Kiffani Lexy 221 12 13
## 222 Kiffani Maya 222 12 14
## 223 Kiffani Mckenna 223 12 15
## 224 Kiffani Megan 224 12 16
## 225 Kiffani Michaela 225 12 17
## 226 Kiffani Richard 226 12 18
## 227 Kiffani Wyatt 227 12 19
## 228 Kiffani Zackary 228 12 20
## 229 Lexy Alexander 229 13 1
## 230 Lexy Andrey 230 13 2
## 231 Lexy Breanna 231 13 3
## 232 Lexy Charles 232 13 4
## 233 Lexy Colton 233 13 5
## 234 Lexy Crystal 234 13 6
## 235 Lexy Derek 235 13 7
## 236 Lexy Francesca 236 13 8
## 237 Lexy Justin 237 13 9
## 238 Lexy Kayla 238 13 10
## 239 Lexy Kelsey 239 13 11
## 240 Lexy Kiffani 240 13 12
## 241 Lexy Maya 241 13 14
## 242 Lexy Mckenna 242 13 15
## 243 Lexy Megan 243 13 16
## 244 Lexy Michaela 244 13 17
## 245 Lexy Richard 245 13 18
## 246 Lexy Wyatt 246 13 19
## 247 Lexy Zackary 247 13 20
## 248 Maya Alexander 248 14 1
## 249 Maya Andrey 249 14 2
## 250 Maya Breanna 250 14 3
## 251 Maya Charles 251 14 4
## 252 Maya Colton 252 14 5
## 253 Maya Crystal 253 14 6
## 254 Maya Derek 254 14 7
## 255 Maya Francesca 255 14 8
## 256 Maya Justin 256 14 9
## 257 Maya Kayla 257 14 10
## 258 Maya Kelsey 258 14 11
## 259 Maya Kiffani 259 14 12
## 260 Maya Lexy 260 14 13
## 261 Maya Mckenna 261 14 15
## 262 Maya Megan 262 14 16
## 263 Maya Michaela 263 14 17
## 264 Maya Richard 264 14 18
## 265 Maya Wyatt 265 14 19
## 266 Maya Zackary 266 14 20
## 267 Mckenna Alexander 267 15 1
## 268 Mckenna Andrey 268 15 2
## 269 Mckenna Breanna 269 15 3
## 270 Mckenna Charles 270 15 4
## 271 Mckenna Colton 271 15 5
## 272 Mckenna Crystal 272 15 6
## 273 Mckenna Derek 273 15 7
## 274 Mckenna Francesca 274 15 8
## 275 Mckenna Justin 275 15 9
## 276 Mckenna Kayla 276 15 10
## 277 Mckenna Kelsey 277 15 11
## 278 Mckenna Kiffani 278 15 12
## 279 Mckenna Lexy 279 15 13
## 280 Mckenna Maya 280 15 14
## 281 Mckenna Megan 281 15 16
## 282 Mckenna Michaela 282 15 17
## 283 Mckenna Richard 283 15 18
## 284 Mckenna Wyatt 284 15 19
## 285 Mckenna Zackary 285 15 20
## 286 Megan Alexander 286 16 1
## 287 Megan Andrey 287 16 2
## 288 Megan Breanna 288 16 3
## 289 Megan Charles 289 16 4
## 290 Megan Colton 290 16 5
## 291 Megan Crystal 291 16 6
## 292 Megan Derek 292 16 7
## 293 Megan Francesca 293 16 8
## 294 Megan Justin 294 16 9
## 295 Megan Kayla 295 16 10
## 296 Megan Kelsey 296 16 11
## 297 Megan Kiffani 297 16 12
## 298 Megan Lexy 298 16 13
## 299 Megan Maya 299 16 14
## 300 Megan Mckenna 300 16 15
## 301 Megan Michaela 301 16 17
## 302 Megan Richard 302 16 18
## 303 Megan Wyatt 303 16 19
## 304 Megan Zackary 304 16 20
## 305 Michaela Alexander 305 17 1
## 306 Michaela Andrey 306 17 2
## 307 Michaela Breanna 307 17 3
## 308 Michaela Charles 308 17 4
## 309 Michaela Colton 309 17 5
## 310 Michaela Crystal 310 17 6
## 311 Michaela Derek 311 17 7
## 312 Michaela Francesca 312 17 8
## 313 Michaela Justin 313 17 9
## 314 Michaela Kayla 314 17 10
## 315 Michaela Kelsey 315 17 11
## 316 Michaela Kiffani 316 17 12
## 317 Michaela Lexy 317 17 13
## 318 Michaela Maya 318 17 14
## 319 Michaela Mckenna 319 17 15
## 320 Michaela Megan 320 17 16
## 321 Michaela Richard 321 17 18
## 322 Michaela Wyatt 322 17 19
## 323 Michaela Zackary 323 17 20
## 324 Richard Alexander 324 18 1
## 325 Richard Andrey 325 18 2
## 326 Richard Breanna 326 18 3
## 327 Richard Charles 327 18 4
## 328 Richard Colton 328 18 5
## 329 Richard Crystal 329 18 6
## 330 Richard Derek 330 18 7
## 331 Richard Francesca 331 18 8
## 332 Richard Justin 332 18 9
## 333 Richard Kayla 333 18 10
## 334 Richard Kelsey 334 18 11
## 335 Richard Kiffani 335 18 12
## 336 Richard Lexy 336 18 13
## 337 Richard Maya 337 18 14
## 338 Richard Mckenna 338 18 15
## 339 Richard Megan 339 18 16
## 340 Richard Michaela 340 18 17
## 341 Richard Wyatt 341 18 19
## 342 Richard Zackary 342 18 20
## 343 Wyatt Alexander 343 19 1
## 344 Wyatt Andrey 344 19 2
## 345 Wyatt Breanna 345 19 3
## 346 Wyatt Charles 346 19 4
## 347 Wyatt Colton 347 19 5
## 348 Wyatt Crystal 348 19 6
## 349 Wyatt Derek 349 19 7
## 350 Wyatt Francesca 350 19 8
## 351 Wyatt Justin 351 19 9
## 352 Wyatt Kayla 352 19 10
## 353 Wyatt Kelsey 353 19 11
## 354 Wyatt Kiffani 354 19 12
## 355 Wyatt Lexy 355 19 13
## 356 Wyatt Maya 356 19 14
## 357 Wyatt Mckenna 357 19 15
## 358 Wyatt Megan 358 19 16
## 359 Wyatt Michaela 359 19 17
## 360 Wyatt Richard 360 19 18
## 361 Wyatt Zackary 361 19 20
## 362 Zackary Alexander 362 20 1
## 363 Zackary Andrey 363 20 2
## 364 Zackary Breanna 364 20 3
## 365 Zackary Charles 365 20 4
## 366 Zackary Colton 366 20 5
## 367 Zackary Crystal 367 20 6
## 368 Zackary Derek 368 20 7
## 369 Zackary Francesca 369 20 8
## 370 Zackary Justin 370 20 9
## 371 Zackary Kayla 371 20 10
## 372 Zackary Kelsey 372 20 11
## 373 Zackary Kiffani 373 20 12
## 374 Zackary Lexy 374 20 13
## 375 Zackary Maya 375 20 14
## 376 Zackary Mckenna 376 20 15
## 377 Zackary Megan 377 20 16
## 378 Zackary Michaela 378 20 17
## 379 Zackary Richard 379 20 18
## 380 Zackary Wyatt 380 20 19Note that any observed event involving actor pairs outside
manual.riskset will be automatically added to the risk set,
ensuring that the likelihood is well-defined for all observed
events.
The grid visualization below illustrates the manual risk set at time
\(t_3\), with dyads involving
"Richard" and "Francesca" excluded (white
tiles).
Here, tiles for dyads involving "Richard" and
"Francesca" are white (excluded); the risk set contains
only dyads among "Colton", "Kayla", and
"Lexy".
For undirected networks, the same manual.riskset logic
applies but the grid visualization focuses on the lower triangular,
since actor pairs are sorted alphanumerically.
The processed risk set
After processing, the remify object exposes the risk set
through riskset_info (when
attach_riskset = TRUE, which is the default). The key field
is riskset_info$included, which contains the decoded table
of at-risk dyads:
reh_active <- remify(
edgelist = randomREH$edgelist,
directed = TRUE,
ordinal = FALSE,
model = "tie",
actors = randomREH$actors,
riskset = "active",
origin = randomREH$origin,
riskset_decode = "labels"
)
head(reh_active$riskset_info$included)## actor1 actor2 dyadID actor1ID actor2ID
## 1 Alexander Andrey 1 1 2
## 2 Alexander Breanna 2 1 3
## 3 Alexander Charles 3 1 4
## 4 Alexander Colton 4 1 5
## 5 Alexander Crystal 5 1 6
## 6 Alexander Derek 6 1 7## [1] 380For the full risk set, the dyad map is accessible directly via
reh$index$dyad_map. For active or manual risk sets, it is
in reh$index$dyad_map_active.
The per-event dyad IDs (mapping each observed event to its position
in the risk set) are stored in reh$ids$dyad:
## [1] 182 84 203 4 353 117 334 254 102 56For actor-oriented modeling with an active or manual risk set, sender
and receiver risk sets are stored in reh$sender_riskset and
reh$receiver_riskset respectively, and
reh$index$sender_map provides the mapping between sender
IDs and actor names.