Configure Coarse Classes

To start assessment and adjustment of results of fine and coarse classes generation, it is required to press Activate button that is located to the right from Input status field. To enable the editing mode, Edit binning button is used. To save introduced changes, it is required to press Apply button if it is not required to save changed settings, press Cancell button.

The following commands are available on the toolbar:

• Hide side bar: press this button to hide the bar with a list of input columns, press it again to return the bar.
• Edit binning: the fields used for editing of the coarse classes parameters become available by pressing this button.
• Merge with previous: enables to merge the coarse class selected in the chart or detailing table with the previous (located more to the left) class (the button is available in the binning editing mode, namely, when button has been pressed).
• Merge with next: enables to merge the coarse class selected in the chart or detailing table with the next (located more to the right) class (the button is available in the binning editing mode).
• Show coarse class table: the coarse class area will be represented in the table form (it can be also called by pressing Alt+One).
• Show coarse class charts: the coarse class area will be represented in the chart form (Alt+Two).

All listed commands are also available in the contex menu (the right mouse button). Besides, the following commands are also available in the context menu:

• Bound split: enables to set the new additional bound and, thus, increase the coarse classes count by 1 (this operation is opposed to Merge with next/Merge with previous commands).
• Lock selected: enables to change the field status to Locked (thus, the field will not be used during retraining). The command is also available by pressing button to the left of each input column.
• Unlock selected: enables to change the field status to Unlocked (thus, the field can be used in the process of the coarse classes generation while retraining). The command is also available by pressing button to the left of each input column.

The following parameters can be edited:

• Coarse сlasses:

  • Minimum weight, %. Class weight means the ratio of observations count for which the input indicator value is included into this class to the total count of the initial data set observations. 5% is set be default. It is not allowed to generate the classes with the weight that is less than the set one. The low class weight denotes its low significance and necessity of union with another class.
  • Maximum count. The highest allowable count of classes generated by the handler for a column. Value 5 is set by default. This value can be changed. The high count of classes causes decrease of their weight, whereas the low one brings about decrease of information value.
  • Set... shows dependence IV on classes count and enables to select their number to set the maximum count. Upon pressing Calculate button, the parameters are recalculated according to the selected count of classes.

• Optimization:

  • Uniformity. It defines the method of partitioning of the input indicator variation range into classes (bins). This parameter is equal to 0 by default, and it can vary from 0 to 1. The value that is equal to 0 provides for such partitioning into classes that maximizes the information value. When the parameter value is equal to 1, the algorithm generates classes in such a way to include roughly the same number of observations into each of them. Thus, in the first case, the indicator significance is increased, whereas in the second case, interpretability of the coarse classes is increased.

Any introduced changes are immediately displayed in the chart and detailing table.

It is possible to sort and filter records in the List of input columns. For this purpose, it is required to open pop-down menu by pressing button that appears when cursor hovering over Column or IV caption.