We are releasing an update for ROIstat! This release has a number of new features and fixes.
You can download it here:
(You can read the complete ROIstat Changelog here)
- New packages needed
- To support Bayesian confidence interval calculation, there are a two new packages that will be part of ROIstat's requirements.
install.packages("bayestestR")install.packages("bayesboot")
- A number of packages have had an update, so you should go ahead and update all of them. You can do this in RStudio by clicking Tools | Check for Package Updates and update whatever needs updating. If you use RGui, you can click on Packages | Update packages... to do the same thing.
- To support Bayesian confidence interval calculation, there are a two new packages that will be part of ROIstat's requirements.
- New Features
- Added a feature for histograms when plotting a single column to put the name of the plot style as the title
- Added a feature to frequency polygons to allow the polygons to all be plotted on the same graph when "Display on same axis" is selected
- Added a feature to get the Bayesian credible interval, either HDI or ETI along with an explanation of what each is and is not
- Added an average and overall kappa calculation for discrete gauge analysis
- Added repeated measures for discrete gauge analysis
- Added calculation for kappa control chart limits
- Added calculation for kappa(critical) for discrete gauge capability
- Added a proportion measurement error calculation to the MSA for continous gauges
- UI improvement
- Cleaned up the labeling on kappa agreement under correlation and association
- Math improvement
- Changed missclassification calculations to use integration instead of Monte Carlo to speed it up
- Bug fixes
- Fixed a bug that would cause an error if column lengths were different when calculating confidence intervals
- Fixed a bug that could re-order a boxplot with outliers or wild outliers
- Fixed a bug that would change the ID of one of the data entry boxes on nonparametric tests, changing results
- Fixed a bug that reported incorrect numbers for Correlation and Association | Cohen's kappa due to a change in lolcat
- Fixed a minor bug in discrete analysis that wouldn't add an asterisk if the p-value was below 1-confidence
- Fixed a minor bug that would not give you the mouseover information in the scatter plot if you had changed the names of the variables since importing the data
- Fixed a bug that would fail to calculate intervals if the factor was text instead of a number
- Known issues:
- datamods no longer will load .dat files with only one column - this is a limitation in the packages it uses. A fix is in the works. A workaround is to save that as a .txt file or just cut and paste the data.
- If you run the app from RStudio and you have clicked on a line number to add a breakpoint, it will take a long time to load the app - this is due to RStudio functions that take 10+ seconds to load. Once they are loaded, it does not affect the app. Consider loading via RGui if this is an annoyance since it loads immediately there. This does not happen if you avoid adding a breakpoint that way.
- Lin-Mudholkar test fails due to breaking the function when the correlation is 1 (lolcat issue)
- No button on MVP(E) or Mann-Whitney-U test to stop early - may be added in the future
- The density plot x-axis for the exponential will start at zero even if the minimum is above that. This seems to be built into ggh4x