Plot memory-network communities

type = "physical" (default)

every community as a pebble around the states it holds, through plot.net_hg() on the community hypergraph. A state shared by several communities lies inside each of their pebbles. Every state is a circle whose area follows its flow, with a triangle pointing at the state the walk most often moves to next, and every community has a title box with its flow.

type = "network"

the network of states as a transition network analysis plot through cograph::overlay_communities(): every arrow is a transition with its probability (transitions below 0.05 hidden), every state a circle whose area follows its flow, coloured (Okabe-Ito) by the community that holds most of its flow, and every community a blob around its states, so a shared state lies in two blobs.

type = "states"

the network of memory nodes drawn in the same way, laid out by the walk's transitions between them so that the nodes of a community sit together. Memory nodes with no flow are left out unless show_zero_flow = TRUE, and the title says how many. With more than 20 nodes the probabilities are not printed on the arrows.

Memory nodes with no flow (reached only by random jumps, see hg_communities()) carry no flow of their states and are never drawn in the physical view. The tables returned by hg_get.hypergraphs_memory_communities() are unaffected.

Usage

# S3 method for class 'hypergraphs_memory_communities'
plot(x, type = c("physical", "network", "states"), show_zero_flow = FALSE, ...)

Arguments

x

A hypergraphs_memory_communities object.

type

"physical" (default), "network" or "states".

show_zero_flow

Draw the memory nodes with no flow in the type = "states" view? Default FALSE.

...

For "physical", passed to plot.net_hg() (e.g. seed, label_size, arrow_style = "outside", layout); for "network" and "states", passed to cograph::overlay_communities() and on to cograph::splot() (e.g. layout, edge_label_style, threshold, blob_alpha). They override the defaults set here.

Value

For "physical", a ggplot object (print it to draw). For "network" and "states", x, invisibly (cograph draws with base graphics).

Conditions

hypergraphs_bad_input from plot.net_hg() for arguments passed through ... that it rejects.

References

Saqr, M., López-Pernas, S., Törmänen, T., Kaliisa, R., Misiejuk, K., & Tikka, S. (2025). Transition network analysis: A novel framework for modeling, visualizing, and identifying the temporal patterns of learners and learning processes. Proceedings of the 15th International Learning Analytics and Knowledge Conference (LAK '25), 351-361. doi:10.1145/3706468.3706513

Examples

seqs <- list(c("a", "h", "b", "a", "h", "b", "a"),
             c("c", "h", "d", "c", "h", "d", "c"))
comm <- hg_communities(hon(seqs, max_order = 2L), trials = 2L)
plot(comm)

plot(comm, type = "network")

plot(comm, type = "states")