ip_fit: Iterative proportional fitting to scale seed values to...
In ryantimpe/ipfitr: Iterative Proportion Fitting for the Tidyverse

Description Usage Arguments Value Examples

Iterative proportional fitting to scale seed values to multiple target subtotals.

ip_fit(datatable, targets, datatable.value.name = "value",
  target.value.names = "value", max.error = 0.01,
  max.iterations = 25, freeze_cells = NULL,
  freeze_cells.value.name = "value", freeze_slice = NULL,
  freeze_slice.value.names = "value", minmax_cells = NULL,
  minmax_cells.value.names = c("value_min", "value_max"),
  minmax_slice = NULL, minmax_slice.value.names = c("value_min",
  "value_max"), minmax.smash.param = 1/3, growth_targets = NULL,
  save.tars = FALSE, show.messages = TRUE)

`datatable`	A data frame of values to be scaled to targets.
`targets`	A list of data frames containing subtotal targets for the `datatable`. All totals in each target should be equal.
`target.value.names`	The names of the series in `targets` containing subtotals to scale. Can be string or array of strings.
`max.error`	The maximum total absolute difference allowed between final scaled values and targets. Iterative scaling will complete once the error is below this threshold or `max.iterations` occur.
`max.iterations`	The maximum number of iterations of scaling. Iterative scaling with end once this value is reached, even if the error is above `max.error`.
`freeze_cells`	Optional data frame of values with same series columns as `datatable`, specifying exact values to hit in the scaling. Any rows or values not listed, or `NA`s, will be scaled as normal.
`freeze_cells.value.name`	The name of the series of iced values in `freeze_cells`.
`freeze_slice`	Optional data frame containing subtotal targets for the `datatable`. Unlike `targets`, these data frames can be subsets only containing subtotals for one or more rows. Any rows or values not listed, or `NA`s, will be scaled as normal. Using `freeze_slice` for partial targets will increase the number of required iterations for scaling. This may require the user to increase the value of `max.iterations`. Separate elements of a dimension within a single cell using " + " to scale subtotals.
`minmax_cells`	Optional data frame of values with same series columns as `datatable`, specifying bounded values to hit in the scaling. Provide minimumn and maximum values for a cell to be scaled. Any rows or values not listed, or `NA`s, will be scaled as normal.
`minmax_cells.value.names`	An array of length 2 of the names of the minimum and maximum values in `minmax_cells`.
`minmax_slice`	Optional data frame containing subtotal targets for the `datatable`, specifying bounded values to hit in the scaling. Provide minimumn and maximum values for a slice of the data frame to be scaled. Unlike `targets`, these data frames can be subsets only containing subtotals for one or more rows. Any rows or values not listed, or `NA`s, will be scaled as normal. Using `minmax_slice` for partial targets will increase the number of required iterations for scaling. This may require the user to increase the value of `max.iterations`.
`minmax.smash.param`	Numeric value of 0 < x < 1. Following an out-of-bounds occurence for `minmax_cells`, the `minmax.smash.param` is the additional value added to the scaled value to bring it back into bounds. Values close to 0 bind the value to the violated bound, while close to 1 bind the value to the other bound. Values closer to0 will require more iterations to complete.
`growth_targets`	A data frame of cell or slice targets with growth rates over listed dimensions. See `ipfitr::ip_growth_transform` for detail.
`datatable.value.names`	The name of the series in `datatable` containing the values to be scaled.
`freeze_slice.value.name`	The name or names of the series of iced values in `freeze_slice`.
`minmax_slice.value.name`	The name or names of the series of the minimum and maximum values in `minmax_slice`.

A dataframe with the same dimensionality as datatable, with all values scaled to the subtotals specified in each data frame in targets.

tar1 <- data.frame(x = letters[1:2], value = c(50, 50))
tar2 <- data.frame(y = letters[3:5], value = c(20, 40, 40))
tar3 <- data.frame(z = letters[6:10], value = c(10, 20, 30, 40, 10))

tar.list <- list(tar1, tar2, tar3)
df <- ip_create_seed(tar.list) %>% ip_fit(tar.list)