Build and Apply QSAR/QSPR Model — Bayes Classification — Advanced Options Pane

Set advanced options for Bayes classification, including Kullback-Leibler acceptance criteria for fingerprint data, and the smoothing coefficient. To open this pane, select Bayes Classification in the Method option menu in the Build Task, and click the Advanced Options link in the Bayes Classification Settings section.

Bayes Classification — Advanced Options Pane Features

Smoothing coefficient text box

Specify the smoothing coefficient. Values are normally given in the range 0 to 1.

Kullback-Leibler acceptance criteria section

Set options specific to fingerprints. Only available if the Fingerprints options in the Build Task is used. You can choose to either select the top N structures or by distance cutoff.

Keep bits in top N% of structures

Specify the percentage of structures of which the fingerprint bits are kept during training, as ranked by the Kullback-Leibler significance.

Apply distance cutoffs options

Select this option to use distance cutoffs during training.

Positive cutoff option and text box

Select this option to apply a distance cutoff to positively correlated features, and specify the cutoff in the text box. The cutoff is only applied if the Apply distance cutoffs option is selected.

Negative cutoff option and text box

Select this option to apply a distance cutoff to negatively correlated features, and specify the cutoff in the text box. The cutoff is only applied if the Apply distance cutoffs option is selected.