问题

我想将 `shap.plots.bar`（https://github.com/slundberg/shap）图形添加到子图中。类似于这样...

```python
fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(8,20))
for (X, y) in [(x1, y1), (x2, y2)]:
  model = xgboost.XGBRegressor().fit(X, y)
  explainer = shap.Explainer(model, check_additivity=False)
  shap_values = explainer(X, check_additivity=False)
  shap.plots.bar(shap_values, max_display=6, show=False) # ax=ax ?? 
plt.show()

然而，对于 shap.plots.bar，ax 未定义，不像其他绘图方法，如 shap.dependence_plot(..., ax=ax[0, 0], show=False)。是否有办法将多个条形图添加到子图中？

<details>
<summary>英文:</summary>

I&#39;d like to add the `shap.plots.bar` (https://github.com/slundberg/shap) figure to a subplot. Something like this...

```python
fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(8,20))
for (X, y) in [(x1, y1), (x2, y2)]:
  model = xgboost.XGBRegressor().fit(X, y)
  explainer = shap.Explainer(model, check_additivity=False)
  shap_values = explainer(X, check_additivity=False)
  shap.plots.bar(shap_values, max_display=6, show=False) # ax=ax ?? 
plt.show()

However, ax is undefined for shap.plots.bar, unlike some other plotting methods such as shap.dependence_plot(..., ax=ax[0, 0], show=False). Is there a way to add many bar plots to a subplot?

答案1

得分: 1

查看源代码，该函数不会创建自己的图形。因此，您可以创建一个图形，然后使用plt.sca将所需的坐标轴设置为当前坐标轴。

以下是使用文档中的条形图示例代码来执行此操作的方法。

import xgboost
import shap
import matplotlib.pyplot as plt

X, y = shap.datasets.adult()
model = xgboost.XGBClassifier().fit(X, y)

explainer = shap.Explainer(model, X)
shap_values = explainer(X)

fig, (ax1, ax2) = plt.subplots(2, 1)
plt.sca(ax2)
shap.plots.bar(shap_values)
fig.tight_layout()
fig.show()

如果您确实想要使用ax参数，您需要编辑源代码以添加该选项。

import xgboost
import shap
import matplotlib.pyplot as plt

X, y = shap.datasets.adult()
model = xgboost.XGBClassifier().fit(X, y)

explainer = shap.Explainer(model, X)
shap_values = explainer(X)


fig, (ax1, ax2) = plt.subplots(2, 1)
plt.sca(ax2)
shap.plots.bar(shap_values)
fig.tight_layout()
fig.show()