MOE XGBoost Model Learner

Calculates a XGBoost QSAR Model. The activity data con be in linear or binary form
The Model Evaluate node can be used to calculate predictions.

Options

Activity Field
The database field containing the data to be fitted
Weight Field
The field used to weight individual entries of the database. These weights should be non-negative; negative weights will be set to zero for the calculation. If the Weight Field is set to (NONE), then all entries will have a weight of 1.
Descriptors
Select all numerical descriptor fields that should be included in the model
Step Size
Step size shrinkage, a value between 0 and 1 that controls the contribution of each tree to the final ensemble prediction; also known as learning rate or eta. If zero then its optimal value will be found.
Max. Tree Depth
Maximum tree depth. If zero then its optimal value will be found.
Gamma
Minimum loss reduction required to make a further node partition. If zero then its optimal value will be found.
Min. Child Weights
Minimum sum of instance weights. If activity field is not binary, this specifies the minimum number of molecules in every tree node. If zero then its optimal value will be found.
Fraction of Descriptors
The fraction of descriptors, randomly selected, that will be used to construct a tree. The value must be between 0 and 1. If zero then its optimal value will be found.
Fraction of Molecules
Subsample ratio of the training instances; specifies the fraction of molecules, randomly selected, that will be used to construct a tree. The value must be between 0 and 1. If zero then its optimal value will be found.
Max. Number of Trees
Maximum number of trees. The number of trees is found by early stopping, but can be limited by this value. The minimum value is 1.
Write Report to File
Write model report to a text file. If off the report will only be included in the summary view! If on the filename is also exported as flow variable MoeModelReport
Report Filename
Select txt file to write to

Input Ports

Icon
Molecules with Activity Data

Output Ports

Icon
MOE XGBoost Model

Popular Predecessors

  • No recommendations found

Popular Successors

  • No recommendations found

Views

MOE XGBoost Model Learner
Summary

Workflows

  • No workflows found

Links

Developers

You want to see the source code for this node? Click the following button and we’ll use our super-powers to find it for you.