estimate_n: Estimate count sum

In richardjtelford/count_checker: Check assemblage count sums and percent

Description

Estimate count sum from percent data

Usage

estimate_n(
  x,
  percent_col = "percent",
  taxon_col = "taxon",
  ID_cols,
  digits = 2,
  nmin = 1L,
  nmax = 1000L
)

Arguments

x	data.frame of percent data in long format
percent_col	character giving name of column with percent data
taxon_col	character giving name of column with taxon names
ID_cols	character vector giving name of one or more columns with sample IDs
digits	numeric; how many decimal places are the percent reported to
nmin	integer; possible count sum to start direct search at
nmax	integer; possible count sum to end direct search at

Details

Estimates the count sum with two methods. The first estimate is 100/p_min where p_min is the minimum percent. The second estimate is the value N within the range nmin:nmax for which all p/100 * N is an integer for all percent p. The proportion of taxa for which this is an integer is given by the score. A warning is given if the score is below 1. This can happen nmax is less than the count sum, or if the percent have been incorrectly calculated (perhaps an error in rounding) or percent are calculated from different count sums for different taxa (eg. upland trees shrubs and herbs vs aquatics).

@return A nested tibble #'

• ID_cols - one or more user specified columns identifying the assemblages

• data - nested raw data from each assemblage

• direct_search - nested raw output of the direct_search method

• direct_search_est - nested with the value(s) of est_n_direct with the highest score. If the score is <1 a warning is given. This can contain more than row, especially if the precision of the percent data is low and the estimated count sum is high.

• minpc - the minimum percentage for each assemblage

• est_n_minpc estimate count sum for the minimum percent method

• est_min_minpc - lowest possible count sum given precision of the percentage data

• est_max_minpc - ditto for highest possible count sum

• n_taxa - number of taxa

References

Telford (2019) Tools for identifying unexpectedly low microfossil count sums. Preprint.

Examples

require("dplyr")
data(last_chance)
last_chance0  %>% 
  tidyr::pivot_longer(cols = -c(age_calBP, totcaps), names_to = "taxon", values_to = "percent") %>% 
  estimate_n(digits = 2, ID_cols = c("age_calBP", "totcaps"))

richardjtelford/count_checker documentation built on Jan. 31, 2022, 7:12 p.m.