[Retracted] Marketing System Construction and Risk Analysis Based on Random Forest of Machine Learning

Table 4

Important parameters of the random forest model.

Parameter	Explain	Impact on model performance
N_estimators	Number of base evaluators. The default is 100	It does not affect the complexity of a single tree, but too small is easy to underfit, and too large is easy to overfit
Max_depth	Maximum tree depth, default maximum depth	Maximum depth, i.e., maximum complexity, reduces max_depth to reduce model complexity
Min_samples_leaf	The minimum number of leaf nodes is 1 by default	The default value is the highest complexity, increasing min_samples_leaf to reduce model complexity
Min_samples_split	The minimum number of samples required for internal node division is 2 by default	The default value is the highest complexity, increasing min_samples_split to reduce model complexity
Max_feature	The maximum number of features is auto by default, and the total number of features is squared	The default value is the intermediate complexity, reducing max_feature to reduce the complexity of the model; on the contrary, it increases the complexity of the model