NEWS
dmai 1.0.0
Changes visible to users
dmai() keeps its name, its arguments (.data, method, logbase)
and its result (a table with the date and DMAI); new arguments are
appended: date, x, r (column names; by default Date, x1,
x2, … and r1, r2, … as before), user_cost (user costs given
directly), benchmark (a benchmark rate instead of the largest
component rate, with the upper envelope), percent, base and
as_dt.
- The table also has the index level (
index, equal to base at the
first date) and value (the aggregate in the units of the
components). It is a data.table (version 0.5.0 returned a grouped
tibble).
as_dt = FALSE returns the full result, of class
c("dmai_index", "dmai_result", "list"): the index, the user costs
and shares of every component, the benchmark rate, the dual user-cost
price index, the Divisia second moments and the simple sum, with
print(), summary(), plot(), ggplot2::autoplot() and
as.data.table() methods; the printed result ends with the sources of
the method.
- Every function that takes data or a result is an S3 generic with a
.default method.
method has a default ("Barnett"): in version 0.5.0 a call without
method stopped with an error.
Corrections
- The growth
DMAI of the Barnett method used logarithms to base 10 by
default (logbase = NULL), which gives 0.434 times the growth rate;
Barnett’s Divisia index uses natural logarithms. The default is now
the natural logarithm, so DMAI is the growth in percent;
logbase = 10 reproduces version 0.5.0 exactly.
New
- Methods
"Fisher" (Fisher ideal index), "SimpleSum" and "CE"
(currency equivalent index of Rotemberg, Driscoll and Poterba 1995) in
dmai().
dmai_compare(): every method on the same data, with average growth
per period and per year.
dmai_user_cost(): user costs and expenditure shares of the
components.
dmai_moments(): Divisia means and variances of the growth of the
components and of their user costs (Barnett and Serletis 1990).
- The dual user-cost price index by factor reversal.
dmai_interpret(), dmai_cite(), dmai_app(): a plain-language
interpretation of every result, the references of every method and a
built-in shiny application; dmai_theme(), dmai_save() (with WebP),
dmai_notation(), dmai_reference().
- Data:
money_sim (simulated monthly monetary components, 2005
to 2024) and cpi_ch19 (the artificial data of the Consumer Price
Index Manual, whose chained Laspeyres, Paasche, Fisher and Tornqvist
indexes the package reproduces).
- Quarto vignettes in HTML and PDF; complete help pages with the
formulas; tests against version 0.5.0, hand computations and the
published values.
Internal
- dplyr, magrittr, stringr, tibble and tidyr are no longer imported; the
computations are matrix operations and
data.table.
dmai 0.5.0 (2024-03-23)
- Updated to pass CRAN tests
dmai 0.4.0 (2019-05-18)
Bug
- Fixed the bug in logbase argument
dmai 0.3.0 (2019-04-11)
New Features
dmai 0.1.0 (2018-11-02)
New Features