book
Article ID: KB0081820
calendar_today
Updated On:
Description
When using a model built by a training dataset to make predictions on a test dataset, you get missing value/ incomplete cases for the predictions on the test dataset. If cases in the test dataset have missing values on those variables selected as predictors in the trained model, they will not get predicted results from the trained model.
Issue/Introduction
Why the predicted value for testing sample is missing?
Resolution
In many data mining modules of Statistica, missing data in the training set can be handled by the algorithm.
Here are a few examples of how missing value is handled in some Statistica modules:
Boosted Trees for Regression and Classification Overview (Stochastic Gradient Boosting) - Basic Ideas
Random Forests Introductory Overview and Technical Notes