Neural Networks

Neural networks are now included as a calibration method alongside forest and rainforest models. In addition to this, users have more control over the type of model used, depending on the model type.

A number of other minor improvements are included. Calibration progress can be tracked. It is also possible to load existing .quant files without having to regenerate them, though this can be overridden by the user. Each model type is now siloed away for quick switching back and forth. And changes to the model design for machine learning models must be committed prior to a run. Finally, plots can be generated with different sizes.

Rainforest

The software has been updated to include random forest regressions in addition to traditional Lucas-Tooth models. There are two types of forest regression used:

• Forest: run using user-selected lines as variables
• Rainforest: run using individual detector channel counts as variables

In addition to this, custom lines for non-elemental quantification can be used. This requires definition initially, but in rainforest models, the calibration will determine the best energies to use.

Lastly, users can add notes to their calibrations to document changes and other details.

CloudCal

CloudCal has been in beta for almost a year, and is now ready for release. You don't even need to install anything - you can run it from your webrowser at www.xrf.guru.

If you do want to run it locally, you can do so as well. First, you will need to download a copy of R appropriate for your computer (Mac, Windows, or Linux):

https://www.r-project.org/

Next, you will need to install a package called 'shiny' to run it locally. You can do so by pasting this line into the R consul when you launch it:

install.packages(c("Rcpp", "shiny"))

It will ask you to choose a download mirror (you can choose anyone, the result is the same). Then, to run the software:

shiny::runGitHub("leedrake5/CloudCal")

The first time it runs, it may take some time to download the supporting software. After that, you should be good to go. If you'd like to download a copy and run it offline, you can instead download it from GitHub (https://github.com/leedrake5/CloudCal) and then run it locally:

shiny::runApp("your/computer/directory/CloudCal")