打印 3D 网格搜索以实现超参数可视化

Question

试图可视化超参数及其结果，我无法将它们绘制在 3d 图中

我尝试构建一个函数为：

PlotGridSearch(grid,xparam,yparam,zlabels):

被称为

gs= GridSearchCV(DecisionTreeClassifier()
                 ,HyperParams
                 , scoring='accuracy'
                 , cv=50).fit(train_data,train_labels)

 PlotGridSearch(gs
                ,'param_max_depth'
                ,'param_max_leaf_nodes'
                ,'mean_test_score')

但我可以使用从 param_max_depth y param_max_leaf_nodes 提取的正确标签将 mean_test_score 列转换为必要的矩阵（二维数组）

有什么建议吗？

Answer 1

有一个关于3D surface matplotlib plot 的不错的官方文档。

# This import registers the 3D projection, but is otherwise unused.
from mpl_toolkits.mplot3d import Axes3D  # noqa: F401 unused import

import matplotlib.pyplot as plt
from matplotlib import cm
from matplotlib.ticker import LinearLocator, FormatStrFormatter
import numpy as np


fig = plt.figure()
ax = fig.gca(projection='3d')

# Make data.
X = np.arange(-5, 5, 0.25)
Y = np.arange(-5, 5, 0.25)
X, Y = np.meshgrid(X, Y)
R = np.sqrt(X**2 + Y**2)
Z = np.sin(R)

# Plot the surface.
surf = ax.plot_surface(X, Y, Z, cmap=cm.coolwarm,
                       linewidth=0, antialiased=False)

# Customize the z axis.
ax.set_zlim(-1.01, 1.01)
ax.zaxis.set_major_locator(LinearLocator(10))
ax.zaxis.set_major_formatter(FormatStrFormatter('%.02f'))

# Add a color bar which maps values to colors.
fig.colorbar(surf, shrink=0.5, aspect=5)

plt.show()

如果你在 Jupyter Notebook 中运行 python，你可以使用%matplotlib notebook 使其交互。

在您的情况下，如果您想获得正确的 X、Y 和 Z，您可以从 gs.cv_results_ 获取值，即：

m = len(param_grid['max_depth'])
n = len(param_grid['max_leaf_nodes'])
X = np.reshape(gs.cv_results_['param_max_depth'].data,[n,m]) # do mind the order of reshape, it might diff
Y = np.reshape(gs.cv_results_['param_max_leaf_nodes'].data,[n,m])
Z = np.reshape(gs.cv_results_['mean_test_score'],[n,m])