Skip to main content

Table 4 Overview of the results of hyperparameter tuning (HT) and parameter evaluation for the models 4 and 5. For the parameters used in HT, refer to Table 1. The parameter evaluation presents the two most relevant parameters for model parsimony, i.e. the total number of groups and the total number of splits

From: Identification of risk factors for involuntary psychiatric hospitalization: using environmental socioeconomic data and methods of machine learning to improve prediction

  Model 4 Model 5
Parameter (HT)
Maximum depth of the tree 7 5
Minimum number of cases per groups 100 75
Minimum number of cases per split 250 200
Minimum impurity decrease 0.0001 0.0001
Parameter (evaluation)
Total number of groups 54 39
Total number of splits 27 18