| Title: | Ports the Workflow of "Resampling Stats" Add-in to R |
|---|---|
| Description: | Resampling Stats (http://www.resample.com) is an add-in for running randomization tests in Excel worksheets. The workflow is (1) to define a statistic of interest that can be calculated from a data table, (2) to randomize rows ad/or columns of a data table to simulate a null hypothesis and (3) and to score the value of the statistic from many randomizations. The relative frequency distribution of the statistic in the simulations is then used to infer the probability of the observed value be generated by the null process (probability of Type I error). This package intends to translate this logic for R for teaching purposes. Keeping the original workflow is favored over performance. |
| Authors: | Paulo Prado [aut, cre], Andr'e Chalom [aut], Alexandre Oliveira [aut] |
| Maintainer: | Paulo Prado <[email protected]> |
| License: | GPL-2 |
| Version: | 0.1.1 |
| Built: | 2025-02-08 04:07:11 UTC |
| Source: | https://github.com/piklprado/rsampling |
Number of Azteca ants recruited by leaf extracts of their host plant , Cecropia trees.
aztecaazteca
A data frame with 21 rows and 3 variables:
plant id, integer
number of recruited ants in the leaf that received drops of smashed new leaves extract
number of recruited ants in the leaf that received drops of smashed old leaves extract
The ant colonies live in the hollow trunk of Cecropia and can detect and expel leaf-chewing insects. To test if this response is more intense in young leaves, drops of extract of smashed young and old leaves were poured in two neighbor leaves of the same plant. After 7 minutes the number of recruited ants in each leaf was recorded.
Kondrat, H. 2012. Estímulos químicos de folhas novas promovem recrutamento eficiente de formigas associadas à embaúba Cecropia glaziovi (Urticaceae). Curso de campo "Ecologia da Mata Atlântica" (G. Machado; P.I. Prado & A.M.Z. Martini, eds.). Universidade de São Paulo, São Paulo. http://ecologia.ib.usp.br/curso/2012/PDF/PI-Hebert.pdf
Functions to run (un)restricted sampling with or without replacement in a dataframe.
within_rows(dataframe, cols = 1:ncol(dataframe), replace = FALSE, FUN = base::sample) within_columns(dataframe, cols = 1:ncol(dataframe), stratum = rep(1, nrow(dataframe)), replace = FALSE, FUN = base::sample) normal_rand(dataframe, cols = 1:ncol(dataframe), stratum = rep(1, nrow(dataframe)), replace = FALSE, FUN = base::sample) rows_as_units(dataframe, stratum = rep(1, nrow(dataframe)), replace = FALSE, length.out = NULL) columns_as_units(dataframe, cols = 1:ncol(dataframe), replace = FALSE, length.out = NULL)within_rows(dataframe, cols = 1:ncol(dataframe), replace = FALSE, FUN = base::sample) within_columns(dataframe, cols = 1:ncol(dataframe), stratum = rep(1, nrow(dataframe)), replace = FALSE, FUN = base::sample) normal_rand(dataframe, cols = 1:ncol(dataframe), stratum = rep(1, nrow(dataframe)), replace = FALSE, FUN = base::sample) rows_as_units(dataframe, stratum = rep(1, nrow(dataframe)), replace = FALSE, length.out = NULL) columns_as_units(dataframe, cols = 1:ncol(dataframe), replace = FALSE, length.out = NULL)
dataframe |
a dataframe with the data to be shuffled or resampled. |
cols |
columns of dataframe that should be selected to be resampled/shuffled. Defaults for all columns. |
replace |
(logical) should the data be permuted (FALSE) or resampled with replacement (TRUE) ? |
FUN |
function used for the sampling procedure. The default is |
stratum |
factor or integer vector that separates data in groups or strata. Randomizations will be performed within each level of the stratum. Needs at least two observations in each level. Default is a single-level stratum. |
length.out |
(integer) specifies the size of the resulting data set.
For columns_as_units, a data.frame with length.out columns will be returned, and for
rows_as_units, a data.frame with length.out rows will be returned.
Note that if length.out is larger than the relevant dimension, |
a dataframe with the same structure of those input in dataframe with values randomized accordingly.
Each function performs as close as possible the corresponding options in Resampling Stats add-in for Excel (www.resample.com) for permutation (shuffling) and sampling with replacement (resampling) values in tabular data:
normal_rand corresponds to the 'normal shuffle' and 'normal resample' option.
For shuffling (replace=FALSE) the data is permuted over all cells of dataframe.
For resampling (replace=TRUE) data from any cell can be sampled and attributed to any other cell.
within_rows and within_columns correspond to the options with the same names.
The randomization is done within each row or column of dataframe.
So for shuffling the values of each row/column are permuted independently and for
resampling the values are sampled independently from each row/column and attributed only
to cells of the row/column they were sampled.
rows_as_units and columns_as_units also correspond to the options with the same names.
Each row or column dataframe is shuffled or resampled as whole.
Only the placement of rows and columns in the dataframe change. The values and their position within each row/column remains the same.
All functions assemble the randomized values in a dataframe
of the same configuration of the original. Columns that
were not selected to be randomized with argument cols are then
bound to the resulting dataframe. The order and names of the rows and columns are preserved, except if length.out
is specified. In this case, the randomized rows/columns may be shifted to the end of the table.
When both stratum and length.out are used, the function will try to keep the proportion of each strata close to the original.
Statistics.com LCC. 2009. Resampling Stats Add-in for Excel User's Guide. http://www.resample.com/content/software/excel/userguide/RSXLHelp.pdf
Plots the distribution of the statistic of interest. Has switches to plot the extreme values and null hypothesis rejection region (also known as critical region).
dplot(dist, svalue, pside = c("Two sided", "Greater", "Lesser"), extreme = TRUE, vline = TRUE, rejection = TRUE, ...)dplot(dist, svalue, pside = c("Two sided", "Greater", "Lesser"), extreme = TRUE, vline = TRUE, rejection = TRUE, ...)
dist |
the statistic distribution, as generated by
|
svalue |
the result of applying the statistic over the original data |
pside |
the alternative hypothesis for the hypothesis testing |
extreme |
logical. should extreme points be highlighted in the plot? |
vline |
logical. should the svalue be displayed as a vertical line? |
rejection |
logical. should the critical region be highlighted? |
... |
further arguments to be passed to |
See the package vignettes for more information about how to interpret this graph
Presence/absence data of vines on Cecropia trees of two morphotypes.
embaubaembauba
A data frame with 152 rows (plants) and 2 variables:
the tree morphotype, factor with two levels
does the tree harbor vines? Logical.
Two morphotypes of Cecropia trees differ in the occupancy by ant colonies. Ants attack and drive out other insects that get to the trees. To test if this protection also affects infestation by vines, trees of similar size of both morphs were sampled and inspected for the presence of vines.
Mello, T.J. 2012. Infestação por lianas e comportamento de poda por formigas em Cecropia (Urticaceae). Curso de campo "Ecologia da Mata Atlântica" (G. Machado; P.I. Prado & A.M.Z. Martini, eds.). Universidade de São Paulo, São Paulo. http://ecologia.ib.usp.br/curso/2012/PDF/PI-Thayna.pdf
Occupancy of Peucetia spiders on parts of an experimental arena covered by leaves with or without trichomes.
peucetiapeucetia
A data frame with 27 rows (trials) and 6 variables:
Is the spider on the part covered by hairy leaves? Logical, for each of 6 successive inspections (time 1, 2, ...
Spiders of the genus Peucetia do not make webs and hunt actively on the vegetation. The data is from an experiment to test if spiders prefer to stay in hairy leaves, that can stick their prey. The spiders were kept in Petri dishes that had half of lower plate covered with hairy leaves. The other half was covered by leaves without trichomes. The placement of each spider was recorded 6 times at each 30 min.
Werneck, R.T. 2010. Lar, viscoso lar. Experimento de seleção de habitat e forrageio de aranhas em plantas com tricomas glandulares. Curso de campo "Ecologia da Mata Atlântica" (G. Machado; P.I. Prado & A.A. Oliveira, eds.). Universidade de São Paulo, São Paulo. http://ecologia.ib.usp.br/curso/2010/pages/pdf/PI/relatorios/rachel.pdf
Occurrences of aphids of the genus Dactynotus on plants of the genus Solidago in Canada
pieloupielou
A dataframe with 10 rows (aphid species of the genus Dactynotus) and 12 columns (plant species of the genus Solidago). Each entry is the number of records of a given aphid species on a plant species.
Data from a field survey by E.C. Pielou in Ontario to exemplify a method to calculate niche overlap and niche width. The niche overlap gauges the overall similarity of the plant ranges used by the aphids. The niche width expresses how diverse is the average diversity of plants used by the aphids.
Pielou, E.C. 1972. Niche width and niche overlap: a method for measuring them. Ecology, 53: 687–692.
Canopy to height ratio and variables of root area in mangrove trees sampled in two soil types.
rhyzophorarhyzophora
A data frame with 24 rows (trees) and 4 variables:
soil type according to instability; factor with two levels (high / medium)
ratio between canopy and trunk area, both in m2, numeric
area covered by aerial roots, numeric (m2)
number of aerial roots, integer
Data from a field practical exercise to test if mangrove trees in more unstable soil allocates more biomass in supporting roots.
Prado, A. et al. 2013. Variações na morfologia de sustentação em Rhizophora mangle (Rizophoraceae) em diferentes condições de inundação do solo. Curso de campo "Ecologia da Mata Atlântica" (G. Machado, P.I. Prado & A.M.Z. Martini eds.). Universidade de São Paulo, São Paulo. http://ecologia.ib.usp.br/curso/2013/pdf/PO4-2.pdf
Repeats resampling/shuffling of dataframes and scores the values returned by user-define function which is applied to each randomized dataframe.
Rsampling(type = c("normal_rand", "rows_as_units", "columns_as_units", "within_rows", "within_columns"), dataframe, statistics, ntrials = 10000, simplify = TRUE, progress = "text", fix.zeroes = FALSE, ...)Rsampling(type = c("normal_rand", "rows_as_units", "columns_as_units", "within_rows", "within_columns"), dataframe, statistics, ntrials = 10000, simplify = TRUE, progress = "text", fix.zeroes = FALSE, ...)
type |
character; the name of the randomization function to be applied to |
dataframe |
a dataframe with the data to be shuffled or resampled. |
statistics |
a function that calculates the statistics of interest from the dataframe. The first argument should be the dataframe with the data and preferably should return a (named) vector, data frame, matrix or array. |
ntrials |
integer; number of randomizations to perform. |
simplify |
logical; should the result be simplified to a vector, matrix or higher dimensional array if possible? |
progress |
which kind of progress bar should be used (currently unimplemented!) |
fix.zeroes |
logical; for normal_rand, within_rows or within_columns, should zeroes in the dataframe
be kept in place? See the help on |
... |
further arguments to be passed to the randomization functions
(e.g., |
a list of objects returned by the function defined by statistics
or a vector, matrix or array when simplify=TRUE and simplification can be done
(see simplify2array).
This function corresponds to Repeat and score in Resampling Stats add-in for Excel
(www.resample.com). The randomization function defined by type is applied ntrials
times on the data provided by dataframe. At each trial the function defined by argument
statistics is applied to the resulting dataframe and the resulting objects are returned.
Statistics.com LCC. 2009. Resampling Stats Add-in for Excel User's Guide. http://www.resample.com/content/software/excel/userguide/RSXLHelp.pdf
Quick plot of paired differences, for exploratory purposes.
splot(p1, p2, highlight = TRUE, col.dif = c("black", "grey"), ...)splot(p1, p2, highlight = TRUE, col.dif = c("black", "grey"), ...)
p1, p2
|
vectors of paired values (numerical vectors) |
highlight |
should positive and negative differences within pairs highlighted with different colors? Logical |
col.dif |
color vector if |
... |
further arguments to be passed to plotting function ( |
This function builds on sample to provide sampling from a vector, but with all
zero entries fixed. This way, zfsample(c(0,1,0,2)) may result in (0,1,0,2) or (0,2,0,1), but the
positions that were initially zero will remain zeroed.
zfsample(x, replace = FALSE)zfsample(x, replace = FALSE)
x |
Either a vector of one or more elements from which to choose, or a positive integer. |
replace |
Should sampling be with replacement? |
a vector of the same length of 'x' with elements drawn from 'x'.
The actual sampling is done by sample, so its help page should be checked
for details on the parameter handling. The parameter 'size' is always passed as length(x),
and 'prob' is not supported.
# Sampling without replacement zfsample(c(0,1,2,0,3,4,0)) # Sampling with replacement zfsample(c(0,1,2,0,3,4,0), replace=TRUE) # With no zeroes, zfsample just calls sample set.seed(42); s1<-sample(c(1,2,3,4,5,6)) set.seed(42); s2<-zfsample(c(1,2,3,4,5,6)) all.equal(s1, s2)# Sampling without replacement zfsample(c(0,1,2,0,3,4,0)) # Sampling with replacement zfsample(c(0,1,2,0,3,4,0), replace=TRUE) # With no zeroes, zfsample just calls sample set.seed(42); s1<-sample(c(1,2,3,4,5,6)) set.seed(42); s2<-zfsample(c(1,2,3,4,5,6)) all.equal(s1, s2)