The hardware and bandwidth for this mirror is donated by METANET, the Webhosting and Full Service-Cloud Provider.
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[@]metanet.ch.

shinyML 1.0.1 (2021-02-21)

New features

• Histogram plots on shiny_h2o and shiny_spark functions now integrate density curves.
• shiny_h2o and shiny_spark functions ensure reproducibility of results when user reproduce the same parameters for a given machine learning model
• shiny_h2o and shiny_spark functions now work with an input dataset that contains a POSIXct column

shinyML 1.0.0 (2020-10-02)

New features

• shiny_h2o and shiny_spark functions have merged into shinyML_regression function: H2O or Spark can now be chosen just using the framework argument.
• A new function shinyML_classification has been implemented to train and test machine learning models for classification tasks : classification results can be viewed through confusion matrix charts in addition to existing available item on old package versions .
• When the framework is set to H2O for shinyML_regression or shinyML_classification function, authorized model families for auto ML searching can be manually specified.
• Two new info cards have been set on the upper part to precise the type of machine learning task (regression or classification) and the dimension of input dataset.
• Autocorrelation plots are now available for numerical variables on Variable summary tab

Breaking changes

• User interface completely changed on shiny apps for both shinyML_regression and shinyML_classification functions : argonDash and argonR shiny API have been used to make user experience even more friendly.
• Both shinyML_regression and shinyML_classification automatically detect if input dataset contains a time-based column: in that case, training and testing dataset splitting is done in order to respect chronology. On the other case, rows are randomly assigned to training or testing dataset according to a splitting percentage parameter.

shinyML 0.2.0 (2019-10-28)

New features

• Informations about cluster memory and number of used CPU(s) are now available on the left side when running shiny_h2o and shiny_h2o functions
• Three new tabs are now available at the top of the shiny_h2o and shiny_h2o dashboards to explore input data set. The Variable Summary tab allows to check types and box plot of each input variable. The Explore dataset tab gives the possibility to understand dependencies by plotting each data variable as a function of another. An overview of all variables dependencies is also available in the Correlation matrix tab.

Breaking changes

• Output tabs like Compare models performances, Feature importance and Table of results are now hidden when no model has been running. It showed a message indicating that output couldn’t be calculated because no model was trained.
• x input parameter of shiny_h2o and shiny_h2o have been removed to give even more simplicity for the user: the dashboard now indicates at the top right of the dashboard which input variable are available to train the model (output variable y is automatically removed from the list).

Bug fixes

• Intercept term button for Generalized linear regression model has changed on both functions due to problem on the UI: the cursor was not at the right position when selected.
• Link button of Generalized linear regression model doesn’t have any effect on the output variable due to omission to take this parameter in account. This issue has been fixed.

shinyML 0.1.1 (2019-08-07)

Bug fixes

• autoML method is now working on shiny_h2o function: the user now just need to set maximum calculation time.

Breaking changes

• Default share_app argument of shiny_h2o and shiny_spark examples have been set to FALSE.

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.