create_convex_hull_structure: create the convex hull associated with points

In skgallagher/EpiCompare: A Pipeline to Analyze and Compare Infectious Disease Epidemics

Description

I believe this will error unless one projects from the simplex to the lower dimensional space

Usage

create_convex_hull_structure(
  data_points,
  data_columns = NULL,
  .lower_simplex_project = TRUE
)

Arguments

data_points	points to create the convex hull structure from
data_columns	columns of data.frame that relate to the point's coordinates in euclidean space. This should be at least 3 columns (else it doesn't really make sense to use this function). The input should look like something like c(S,I,R) or c("S", "I", "R"). If the input is NULL this function will treat this like dplyr::everything().
.lower_simplex_project	boolean, if data points should be projected to a simplex and then to the lower dimensional simplex (for this package, this should always be done)

Value

a convex_hull_structure object, that is a data frame similar to data_points (but with out data_columns columns) and only points that define the convex hull. This object also has an normals attribute which contains a matrix with the first p columns defining "A" and the last defining "b", such that -Ax >= b defines the convex hull.

Examples

library(dplyr)
# this will look like a 2d data frame (but this is becuase pomp_df info is
# on the 3d simplex)
convex_hull_cb <- EpiCompare::pomp_df %>%
  filter(.id <= 10) %>%
  arrange(time) %>% # just to be safe
  select(-time, -H, -cases) %>%
  group_by(.id) %>%
  grab_top_depth_filaments(conf_level = .9) %>%
  create_convex_hull_structure() #data_columns = c(S,I,R)

skgallagher/EpiCompare documentation built on Sept. 14, 2021, 5:45 a.m.