使用 XGBoostplot_tree 绘制 BQML 增强树回归量

发布于 2025-01-15 09:55:59 字数 800 浏览 2 评论 0原文

我在 BQML 中训练了增强树回归器，并导出了模型。请参阅下面生成的文件结构。

我可以使用 XBoost 导入 model.bst 文件并绘制决策树，但图中缺少特征名称。

    # plot decision tree
    from xgboost import XGBRegressor, plot_tree, plot_importance
    import matplotlib.pyplot as plt
    
    %matplotlib inline
    
    plt.style.use('ggplot')
    
    model=XGBRegressor()
    model.load_model('./boostedtree_model/model.bst')
    
    plt.figure(figsize=(50,50))  # set plot size (denoted in inches)
    
    plot_tree(model,fontsize="48",rankdir='LR')
    plt.savefig('boosted_tree.pdf',bbox_inches='tight',dpi=1200)

据我在互联网上找到的信息，BQML 不会生成包含模型和功能名称的 model.json。

有谁知道如何获取 BQML 训练模型的名称？

原文

I trained a boosted tree regressor in BQML, and exported the model. See below the resulting file structure.

I'm able to import the model.bst file with XBoost and plot the decision tree, but the name of the features are missing from the plot.

    # plot decision tree
    from xgboost import XGBRegressor, plot_tree, plot_importance
    import matplotlib.pyplot as plt
    
    %matplotlib inline
    
    plt.style.use('ggplot')
    
    model=XGBRegressor()
    model.load_model('./boostedtree_model/model.bst')
    
    plt.figure(figsize=(50,50))  # set plot size (denoted in inches)
    
    plot_tree(model,fontsize="48",rankdir='LR')
    plt.savefig('boosted_tree.pdf',bbox_inches='tight',dpi=1200)

For what I could find on the internet, BQML do not generate a model.json, which would contain the model and the feature names.

Does anyone knows how to get the names for BQML trained models?

分享到QQ

分享到微博