What's Changed

• Fix random seed error in newer versions of MATLAB.

Full Changelog: v1.0.0-beta.12_hotfix...v1.0.0-beta.13

What's Changed

• NBS-Predict v1.0.0-beta.12-hotfix by @eminSerin in #40

Full Changelog: v1.0.0-beta.12...v1.0.0-beta.12_hotfix

What's Changed

• Dev by @eminSerin in #38
• dev by @eminSerin in #39

Full Changelog: v1.0.0-beta.11...v1.0.0-beta.12

What's Changed

• NBS-Predict v1.0.0-beta.11 by @eminSerin in #36

Full Changelog: v1.0.0-beta.10-hotfix...v1.0.0-beta.11

What's Changed

• • Hotfix for results viewing bug! by @eminSerin in #34

Full Changelog: v1.0.0-beta.10...v1.0.0-beta.10-hotfix

• Simplified and refreshed GUI
• Users can now visualize positive and negative networks and combined networks in regression analyses.
• KFold used during evaluating subnetwork is set to the one the user defines if it is higher than the defined one.
• Fix minor issues in the legacy GUI
• The toolbox will run the legacy GUI if the MATLAB version is old.
• Solve compatibility issues in the GLM function.
• Redesign the Analysis Window with AppDesigner.
• Several bug fixes.

• Drop Decision Tree Regressor and Classifier to avoid possible incompatibilities with confound regression.
• Minor tweaks in parallel processing.

• NBS-Predict is now in v1.0.0-beta.8
• Fix minor bug in feature selection (when confounds are provided in the
  design matrix, but not specified in the contrast vector.).

• NBS-Predict is now in v1.0.0-beta.7
• Trained model is now automatically extracted.
• Add function (NBSPredict_predict) to predict novel data using the trained model.
• Refactor several functions to make them compatible with the novel data prediction function.
• The symmetry of the connectivity matrices is now automatically checked.
• Fix minor GUI issue in the results viewer window.
• Fix edge weight issue when MAD, MSE, or RMSE is chosen.
• Users can also provide a directory for holdout data.
• Refactor load functions to be compatible for loading a single dataset file (NodesxNodesxSubjects).

• NBS-Predict is now in v1.0.0-beta.6
• User can specify the number of CPU cores to use.
• Fix minor bug in Elastic Net function.
• Add a progress bar for loading connectivity matrices on the command window.
• Minor change in loadData function to fasten .csv loading.
• Add balanced accuracy metric to measure accuracy in imbalanced datasets.
• Add permutation testing to compare model performance against null distribution.