Release Notes Version 7 5

From Eigenvector Research Documentation Wiki
Jump to navigation Jump to search

New Features in Solo and PLS_Toolbox

Hierarchical Model Builder

  • New graphical interface to build Hierarchical Models based on Model Selector models.
  • Allows building of complex classification trees, local model selectors, and input/output filters and tests.
  • Branch decisions based on classification models, regression models, decomposition models, or tests on raw input variables.
  • Output model predictions, fixed strings or numerical values, or throw errors for each branch.
  • Drag and drop models and other objects from the Model Cache, workspace, or file system.
  • Compatible with Solo_Predictor for immediate use in on-line applications.

Model Optimizer

  • Added ability to automatically survey over preprocessing settings.
  • Improved support for PLSDA and SVMDA, added support for new ANN models.
  • Improved handling of options and modeling errors.

Artificial Neural Networks

  • Artificial Neural Network (ANN) added to Analysis window and command-line tools. Build ANN models in the same familiar interface as other models.
  • Initial release includes previously released engine/algorithm, next minor release will include an additional, previously unreleased algorithm.

Analysis and Workspace Browser

  • Completely revamped options window allows easier saving and restoring of options.
  • Export and Import models and calibration data directly from the Model Cache to simplify sharing of models and data with others.
  • Easier access to build SVM classification models and SIMCA models on any class set.
  • Improved control over SIMCA classification rules and options.
  • Add Refine menu to Analysis to better organize menu items and provide access to new hierarchical model builder.
  • Add option to browse (see Edit > Options > Workspace Browser Options) to do either concatenation or not when dropping multiple files.

New/Improved Importers and Preprocessing Methods

Plot Controls Improvements

  • Added context-menu options to quickly select all members of a given sample's class
  • Added ability to get summary statistics on selected samples (e.g. get mean of all selected samples)


New Command-line Features and Functions

Misc New Functions

  • ann - Artificial Neural Network regression models.
  • dataset/copy_clipboard - Copy data and labels to clipboard.
  • EGrid - Property grid control based on Jide Property Table.
  • emscorr - Extended Multiplicative Scatter Correction (EMSC).
  • evritree - Wrapper object for working with jtree in Matlab.
  • evritreefcn - Rename of old evritree function.
  • hideaxistext - Hides text objects which are outside the current axis limits.
  • importtool - GUI for designating column/row data types.
  • modelselectorgui - Graphical Interface to build Hierarchical (modelselector) models.
  • preprocessiterator- Create array of preprocessing combinations.
  • sammon - Computes Sammon projection for data or map.
  • textreadr - Reads an ASCII or .XLS file in as a DataSet Object (formally xclreadr).