The hardware and bandwidth for this mirror is donated by dogado GmbH, the Webhosting and Full Service-Cloud Provider. Check out our Wordpress Tutorial.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]dogado.de.
Due to compatibility with dfidx which is used in the
mlogit package for obtaining odds ratio,
the package dependency is raise to R >= 4.2. The test cycle is now
between R4.2 and R4.6.
Coefficient estimates of probit model are now well
aligned for conversion to odds ratio in
odds_summary.
phi coefficients in odds_summary are
now better controlled. The output variables in mdi are
now better aligned.
There has been huge improvements and higher efficiency as
mdpi and plot_mdpi deprecated in favor of
mdi and plot_mdi as the functions can
now handle any multidimensional indices like poverty,
wellbieng, etc, as long as they have
indicators and dimensions.
Exaggerated Nobs in quicksummary function
is corrected.
qpdf is out of suggests in the first check with
R4.6.0.
Deprecation list introduced. It is now possible to specify the number
of digits in quicksummary. The flow of the output in
quicksummary summary is more logical now, i.e., Min,
Q1, Median, Q3, Max, IQR. betareg and
mvProbit are now supported in
odds_summary.
odds_summary added. select tightened in
testthat. devtools:install_github changed to pak::pak.
case_match replaced with recode_values in
treatment_model and vector list used with
case_when in model_factors following the
release of dplyr 1.2.0.
forecst now demoted; it is no more imported
namespace but is suggested as forecast is now
generics. The DynamicFocast is now more efficient
as a result of this change.
Sen Index, by Amartya Sen, is now evaluated along
with MDPI in the mdpi function.
MaximumDate argument in the function
DynamicForecast is deprecated.
A new function relative_likert added to the package.
A new function index_construction added to the package.
In view of the persistent error thrown up in some parts of
model_factors, the strategy of recovery was changed, hence
the 1768 lines of codes was replaced by a more efficient
22 lines of codes which produces enhanced objects and kills
the error. There is a new argument to estimate
Resilience capacity with model_factors. Some
Arguments in quicksummary are deprecated. The
test coverage increased to 84%.
varImp has been demoted and several editing was carried
out for efficiency. Also, augment function has been demoted
from broom package to generics. The test coverage increased
to 76% since more functions are now tested. The number of dependencies
continue to reduce making the package less vulnerable.
quicksummary now have additional output of five types of
mean, which can be useful when working on various modeling tasks.
Some editing on the various functions to make them more readable.
tidyverse now demoted to suggestion from requirement to
reduce the load of dependencies. PRESS restored as
Predictive Residual Sum of Squares.
Relative importance index now part of the objects created in Garrett
Ranking. The data summary and the Garrett ranking are now in one
data.frame. The Remark column in Likert-type
object in quicksummary is discontinued. It is now left to
the user to interpret the results obtained.
Some editing to make codes more readable.
More efficient estimation procedure and better logic in selecting the
options to estimate in Linearsystems function. Some options
will not output because of complexity. Vignette is now included in the
documentation with lots of examples. The test coverage has increased
from 39% to 61%. met.brewer function has been retired and
MetBrewer package was demoted to suggestion rather than
import.
Some editing on the various functions to make them more readable.
Huge improvements and new additions to the package. A lot of
editorial corrections observed in the codes have been carried out. Tree
new functions, mdpi, plot_mdpi and
gender are successfully tested and now operational. The
functions provides for awesome resources for simple, precise and
empirical policy making processes and procedures.
The functions are for computing and visualizing multidimensional
poverty indices. There is also more clarity of the difference between
sex and gender.
roxygen 2 misleading internal keyword demobilised to index all the help pages.
modelsummary’s datasummary_df insistence on tinytable object for some summary data addressed, The datasummary_df was demobilised.
treatment_model introduced for propensity score matching and treatment effects
PRESS suspended in MlMetrics
minor bugs discovered in the garret ranking function modules fixed.
minor bugs fixed probably due to changes in dependencies following new R release.
timeDate dependency in quicksummary is
removed.
Huge improvements and new additions to the package. A lot of
editorial corrections observed in the codes have been carried out. Two
new functions, Model_factors and
garrett_ranking are successfully tested and now
operational. The two functions provides for gaining deeper insights into
the meaning behind Likert-type variables collected from respondents.
Garrett ranking provides the ranks of the observations of the variables
based on the level of seriousness attached to it by the respondents. On
the other hand, Model factors determines and retrieve the latent factors
inherent in such data which now becomes continuous data. The factors or
data frame retrieved from the variables can be used in other analysis
like regression and machine learning.
The two new functions are for factor analysis, essentially,
exploratory factor analysis (EFA), used to unravel the underlying
structure of the observed variables. The analysis also helps to reduce
the complex structure by determining a smaller number of latent factors
that sufficiently represent the variation in the observed variables.
With EFA, no prior knowledge or hypothesis about the number or nature of
the factors is assumed. These are great tools to help tell the story
behind your data. The data used for Model_factors is
prepared using fa.parallel and fa functions in
the psych package.
The Data argument in the
DynamicForecast function has been modified. Rather than the
previous argument were a data frame, containing the two required
vectors, was required, the data argument is now in three vectors
date, series, x. The x argument is new but
optional and would now allow additional variables apart from the two
mandatory ones to be added to the model. The origin
argument which was previously embedded in ... has now being
defaulted to 1970-01-01. Two objects have been added to
the results, i.e., table of estimates and the
plots of the fitted values. The function, which was the
first in the package has now be promoted to stable.
The new version is coming with a huge improvement. The number of functional forms of the linear regression has increased from seven to 14 based on the Taylor Series expansion of the Box-Cox transformations. In addition, two new plots are added to the suite i.e. fitted and effects plots. Also, Tables of marginal effects are now produced for fast inference from the estimated coefficients. This new version also comes with 47 metrics for assessing model performance and carry out model diagnostics. This new version also comes with the capacity of making prediction about each of the functional forms.
There are two new functions coming along with this new version.
corplot for custom plotting of the correlation matrix of
the variables included in the model. estimate_plot is for
plotting the estimated coefficients in their order of significance in
the estimated model. In addition, new plot, visual means
has made it easier to understand the nature of the distribution of the
variables, digest and make inference about the models estimated.
Even with this huge improvements, the principle of one line of code
is maintained. It can be implemented easily once the data has been
cleaned. The structure is
Linearsystems(y, x, mod, limit, Test). The argument
Test is optional, if omitted, the prediction is made with
NULL data.
As always, comments, questions and suggestions are welcome.
quicksummary now returns a data frame rather a
kable Table.
The internal argument digits in Percent is
removed due to observed exaggeration of the final rates
Final formatting of data frames in the Linearsystems
package was removed. Now, the estimates can display in the
console.
A new function data_transform added and is working
perfectly. The addition of this function only makes the package more
versatile for Machine Learning operations and
estimations.
A new function formattedcut added and is working
perfectly. The addition of this function is providing more clarity to
the package. Quite a number of editorial fixes were undertaken and the
codes are working superbly.
Mallow’s Cp successfully added and documentation improved sufficiently. The codes are working better now and the check results are encouraging.
Documentation improved and new functions added to the library. Editorial fixes were made.
A small experimental library with only one function, i.e., DynamicForecast and is perfectly working although the check result looks funny.
NEWS.md file to track changes to the
package.These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.
Health stats visible at Monitor.