• How-To
TreeNet® Classification
Before you start
  • Overview
  • Data considerations
  • Example
  • Example of discover key predictors
  • Example of prediction
Perform the analysis
  • Enter your data
  • Predictor elimination
  • Specify the class weights
  • Specify the interactions
  • Specify the validation method
  • Select the analysis options
  • Select the graphs to display
  • Select the results to display
  • Store statistics
Select an alternative model
  • Select an alternative model from Discover Key Predictors
Tune hyperparameters
  • Select hyperparameter values to evaluate
Add partial dependence plots
  • Select more predictors to plot
Predict new results
  • Predict new results
  • Select the prediction results to display
  • Store prediction statistics
Interpret the results
  • Method table
  • Response information table
  • Model evaluation
  • Optimization of hyperparameters
  • Average negative log-likelihood vs number of trees plot
  • Area under ROC curve vs number of trees plot
  • Misclassification rate vs number of trees plot
  • Model summary table
  • Relative variable importance chart
  • Top 2-way interaction strength tables
  • Confusion matrix
  • Misclassification table
  • Receiver operating characteristic (ROC) curve
  • Gain chart and Lift chart
  • Boxplot of event probabilities
  • Partial dependence plots
  • Prediction table
Methods and formulas
  • Methods
  • Selection of the optimal number of trees
  • Response information
  • Predictor elimination
  • Tune hyperparameters
  • Fits and probabilities
  • Model summary
  • Confusion matrix
  • Misclassification table
  • Top 2-way interaction strength
  • Receiver Operating Characteristic (ROC) curve
  • Gain chart
  • Cumulative lift chart
  • Lift chart
  • Partial dependence plots