Training: Difference between revisions

From ACD Percepta
Jump to navigation Jump to search
No edit summary
Line 6: Line 6:
# In the '''Create New Self-training Library''' window, type in library name and description.
# In the '''Create New Self-training Library''' window, type in library name and description.
# Browse for a folder where you wish to save the new library, and enter the desired name of the *.trn file.
# Browse for a folder where you wish to save the new library, and enter the desired name of the *.trn file.
# The newly added library appears under the respective '''Trainable Libraries''' group and is automatically loaded in the '''Spreadsheet''' workspace.
# The newly added library appears under the respective '''Trainable Libraries''' group and is automatically loaded in the '''[[Spreadsheet View|Spreadsheet]]''' workspace.
<br />
<br />



Revision as of 13:59, 10 July 2012

How to:


How to create a new training Library?

  1. In the Data Source tree, select “New Library” in the appropriate Trainable Libraries group.
  2. In the Create New Self-training Library window, type in library name and description.
  3. Browse for a folder where you wish to save the new library, and enter the desired name of the *.trn file.
  4. The newly added library appears under the respective Trainable Libraries group and is automatically loaded in the Spreadsheet workspace.


How to add new compounds from the main program window to the existing training Libraries?

  1. In the appropriate Prediction Module, click the Train Library Icon icon to open a window with settings for the single structure to be designated. This icon is enabled only when the editable training library is selected (clik on the Train Library Configure Icon icon to do this).
  2. Manipulate the structural formula of the compound using standard ACD/Percepta structure editing tools if needed.
  3. Edit the pre-selected or enter a new name for the compound in the Name text field.
  4. Enter the experimentally measured property value (only numerical values are accepted in this field).
  5. Click OK to add the new compound to the selected Self-training Library.


In this case the compound addition dialog box is pre-filled with the compound structure from the main Prediction Module window Structure Pane and, if found, the name of the compound from the internal ACD/Percepta dictionary. This is meant to speed up the procedure in the situations when the user wants to quickly add currently analyzed compound to the Self-training Libraries. However the dialog box provides all the means necessary for any modifications, i.e. the initial pre-selections are not obligatory and the described procedure can be used to add any new compound to the libraries.


How to import new compounds from files to the existing training Libraries?

  1. In the appropriate Prediction Module, click the Train Library Icon icon to open the Train Library window and select “Multiple”. It allows the user to add multiple compounds to the existing Self-training Libraries in one step. This icon is enabled only when the editable training library is selected (clik on the Train Library Configure Icon icon to do this).
  2. Choose the compound import type - import compounds from files (structure database files, tab-delimited text files).
  3. When a file is selected, choose the fields containing compound names and experimentally measured property values from the drop-down list containing a list of all the fields available in the database file.
  4. A progress report is displayed when import is finished.




FAQ:


Why are two versions of each trainable library in ACD/Percepta?

Each library comes in two identical copies – ‘Read-only’ and ‘Editable’. The user is free to edit the contents of the ‘Editable’ version while no alterations are allowed to the ‘Read-only’ library which can be considered as a backup copy of the original data. Otherwise these Built-in Self-training Libraries have the same functionality – both can be used in calculations or as an initial data source for the creation of user-defined Self-training Libraries.