网页设计html代码教程_电商的推广方式_电商推广和网络推广的策略_网站买卖交易平台

dtreeviz库

• dtreeviz有以下特色：

  1. 利用有颜色的目标类别图例

  2. 叶子大小与该叶子中的样本数成正比

  3. 将≥和<用作边缘标签，看起来更清晰

  4. 决策节点利用堆叠直方图展示特征分布，每个目标类别都会用不同的颜色显示

  5. 在每个节点中各目标类别的样本数都用直方图的形式，这样可以提供更多信息

  dtreeviz同样依赖GraphViz。

• dtreeviz安装

  pip install dtreeviz                  # install dtreeviz for sklearnpip install dtreeviz[xgboost]         # install XGBoost related dependencypip install dtreeviz[pyspark]         # install pyspark related dependencypip install dtreeviz[lightgbm]        # install LightGBM related dependency
  

• 分类树

  """分类树"""from sklearn.datasets import load_iris
  from sklearn import tree
  import graphviz
  from dtreeviz.trees import dtreeviz
  from sklearn.tree import DecisionTreeClassifieriris = load_iris()clf = DecisionTreeClassifier(criterion='gini',max_depth=2)# 演示用，只用2层树
  clf.fit(iris.data,iris.target)viz = dtreeviz(clf            #分类器,iris.data      # X,iris.target    # y,target_name='Iris_variety'    ,feature_names=iris.feature_names   # 特征名称,class_names={0:'setosa',1:'versicolor',2:'virginica'}   # 类别名称,title="Decision Tree Classifier- Iris data set"   # 标题名称,scale=2.4  # 缩放比例，数值越大图越大)viz
  

  

• 回归树

  """回归树"""
  from sklearn.datasets import load_boston
  from sklearn import tree
  import graphviz
  from dtreeviz.trees import dtreeviz
  from sklearn.tree import DecisionTreeRegressorboston = load_boston()reg = DecisionTreeRegressor(criterion='mse',max_depth=2)# 演示用，只用2层树
  reg.fit(boston.data,boston.target)viz = dtreeviz(reg   # 回归器,x_data=boston.data,y_data=boston.target,target_name='price',feature_names=boston.feature_names,title="Decision Tree Regressor - Boston housing",show_node_labels = True,scale=2.4)viz
  

  

  ---