SHAP 值用于多输出回归模型

作者： coryroyce

日期更新： 2021/3/4

创建多输出回归模型

创建数据

导入所需的包

[1]:

import pandas as pd
from keras.layers import Dense
from keras.models import Sequential
from sklearn.datasets import make_regression

创建一个自定义函数，用于生成多输出回归数据。注意：此示例创建了 5 个输出/目标/标签，但该方法可以轻松扩展到任意数量的输出。

[2]:

def get_dataset():
    # Create sample data with sklearn make_regression function
    X, y = make_regression(n_samples=1000, n_features=10, n_informative=7, n_targets=5, random_state=0)

    # Convert the data into Pandas Dataframes for easier maniplution and keeping stored column names
    # Create feature column names
    feature_cols = [
        "feature_01",
        "feature_02",
        "feature_03",
        "feature_04",
        "feature_05",
        "feature_06",
        "feature_07",
        "feature_08",
        "feature_09",
        "feature_10",
    ]

    df_features = pd.DataFrame(data=X, columns=feature_cols)

    # Create lable column names and dataframe
    label_cols = ["labels_01", "labels_02", "labels_03", "labels_04", "labels_05"]

    df_labels = pd.DataFrame(data=y, columns=label_cols)

    return df_features, df_labels

创建模型

创建一个 Tensorflow/Keras Sequential 模型。

[3]:

def get_model(n_inputs, n_outputs):
    model = Sequential()
    model.add(Dense(32, input_dim=n_inputs, kernel_initializer="he_uniform", activation="relu"))
    model.add(Dense(n_outputs, kernel_initializer="he_uniform"))
    model.compile(loss="mae", optimizer="adam")
    return model

训练模型

使用上面定义的函数创建数据集。

[4]:

# Create the datasets
X, y = get_dataset()

# Get the number of inputs and outputs from the dataset
n_inputs, n_outputs = X.shape[1], y.shape[1]

使用上面定义的函数加载模型。

[5]:

model = get_model(n_inputs, n_outputs)

训练模型

[6]:

model.fit(X, y, verbose=0, epochs=100)

[6]:

<tensorflow.python.keras.callbacks.History at 0x7f08e9a7e490>

获取模型评估指标以确认训练良好。

[7]:

model.evaluate(x=X, y=y)

32/32 [==============================] - 0s 826us/step - loss: 15.8952

[7]:

15.895209312438965

模型预测

可以手动输入数据，但在本例中，只需使用特征/X 数据中的任意索引形式。

[8]:

model.predict(X.iloc[0:1, :])

[8]:

array([[ -15.026388,  -64.4412  ,  -75.39472 ,  -70.4628  , -126.55638 ]],
      dtype=float32)

获取 SHAP 值和图表

将 Shapley 值应用于模型。

[9]:

!pip install shap
import shap

# print the JS visualization code to the notebook
shap.initjs()

Collecting shap
  Downloading https://files.pythonhosted.org/packages/b9/f4/c5b95cddae15be80f8e58b25edceca105aa83c0b8c86a1edad24a6af80d3/shap-0.39.0.tar.gz (356kB)
     |████████████████████████████████| 358kB 6.0MB/s
Requirement already satisfied: numpy in /usr/local/lib/python3.7/dist-packages (from shap) (1.19.5)
Requirement already satisfied: scipy in /usr/local/lib/python3.7/dist-packages (from shap) (1.4.1)
Requirement already satisfied: scikit-learn in /usr/local/lib/python3.7/dist-packages (from shap) (0.22.2.post1)
Requirement already satisfied: pandas in /usr/local/lib/python3.7/dist-packages (from shap) (1.1.5)
Requirement already satisfied: tqdm>4.25.0 in /usr/local/lib/python3.7/dist-packages (from shap) (4.41.1)
Collecting slicer==0.0.7
  Downloading https://files.pythonhosted.org/packages/78/c2/b3f55dfdb8af9812fdb9baf70cacf3b9e82e505b2bd4324d588888b81202/slicer-0.0.7-py3-none-any.whl
Requirement already satisfied: numba in /usr/local/lib/python3.7/dist-packages (from shap) (0.51.2)
Requirement already satisfied: cloudpickle in /usr/local/lib/python3.7/dist-packages (from shap) (1.3.0)
Requirement already satisfied: joblib>=0.11 in /usr/local/lib/python3.7/dist-packages (from scikit-learn->shap) (1.0.1)
Requirement already satisfied: python-dateutil>=2.7.3 in /usr/local/lib/python3.7/dist-packages (from pandas->shap) (2.8.1)
Requirement already satisfied: pytz>=2017.2 in /usr/local/lib/python3.7/dist-packages (from pandas->shap) (2018.9)
Requirement already satisfied: setuptools in /usr/local/lib/python3.7/dist-packages (from numba->shap) (54.0.0)
Requirement already satisfied: llvmlite<0.35,>=0.34.0.dev0 in /usr/local/lib/python3.7/dist-packages (from numba->shap) (0.34.0)
Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.7/dist-packages (from python-dateutil>=2.7.3->pandas->shap) (1.15.0)
Building wheels for collected packages: shap
  Building wheel for shap (setup.py) ... done
  Created wheel for shap: filename=shap-0.39.0-cp37-cp37m-linux_x86_64.whl size=491624 sha256=d4d0a19e515d857230caed0cc9bd7ad48017557ad8d72898297455efe78376ea
  Stored in directory: /root/.cache/pip/wheels/15/27/f5/a8ab9da52fd159aae6477b5ede6eaaec69fd130fa0fa59f283
Successfully built shap
Installing collected packages: slicer, shap
Successfully installed shap-0.39.0 slicer-0.0.7

这里我们采用上面训练的 Keras 模型，并解释它为什么对单个样本做出不同的预测。

使用 Kernel Explainer 设置解释器（来自 SHAP 的模型无关解释器方法）。

[10]:

explainer = shap.KernelExplainer(model=model.predict, data=X.head(50), link="identity")

获取单个示例的 Shapley 值。

[11]:

# Set the index of the specific example to explain
X_idx = 0

shap_value_single = explainer.shap_values(X=X.iloc[X_idx : X_idx + 1, :], nsamples=100)





显示单个示例的详细信息




[12]:





X.iloc[X_idx : X_idx + 1, :]









[12]:











  
    

      
      特征_01
      特征_02
      特征_03
      特征_04
      特征_05
      特征_06
      特征_07
      特征_08
      特征_09
      特征_10
    

  
  
    

      0
      -0.093555
      0.417854
      -1.655827
      -2.048833
      -0.258209
      -0.989744
      -0.154596
      -0.338294
      1.503827
      -0.514878
    

  







选择要在其上运行单个解释的标签/输出/目标


注意：下拉菜单可以很容易地通过手动设置要解释的标签上的索引来替换。




[13]:





import ipywidgets as widgets









[14]:





# Create the list of all labels for the drop down list
list_of_labels = y.columns.to_list()

# Create a list of tuples so that the index of the label is what is returned
tuple_of_labels = list(zip(list_of_labels, range(len(list_of_labels))))

# Create a widget for the labels and then display the widget
current_label = widgets.Dropdown(options=tuple_of_labels, value=0, description="Select Label:")

# Display the dropdown list (Note: access index value with 'current_label.value')
current_label



















绘制单个示例和单个标签/输出/目标的力图




[15]:





# print the JS visualization code to the notebook
shap.initjs()

print(f"Current label Shown: {list_of_labels[current_label.value]}")

shap.force_plot(
    base_value=explainer.expected_value[current_label.value],
    shap_values=shap_value_single[current_label.value],
    features=X.iloc[X_idx : X_idx + 1, :],
)



























Current label Shown: labels_01







[15]:










  可视化已省略，未加载 Javascript 库！
你是否在此笔记本中运行了 `initjs()`？如果此笔记本来自其他用户，你还必须信任此笔记本（文件 -> 信任笔记本）。如果你在 github 上查看此笔记本，则 Javascript 已被剥离以确保安全。如果你正在使用 JupyterLab，则此错误是因为尚未编写 JupyterLab 扩展。

 




为特定输出/标签/目标创建汇总图。




[16]:





# Note: We are limiting to the first 50 training examples since it takes time to calculate the full number of sampels
shap_values = explainer.shap_values(X=X.iloc[0:50, :], nsamples=100)

































[17]:





# print the JS visualization code to the notebook
shap.initjs()

print(f"Current Label Shown: {list_of_labels[current_label.value]}\n")

shap.summary_plot(shap_values=shap_values[current_label.value], features=X.iloc[0:50, :])



























Current Label Shown: labels_01













../../../_images/example_notebooks_tabular_examples_model_agnostic_Multioutput_Regression_SHAP_39_2.png





汇总图注释


根据上面的汇总图，我们可以看到特征 01、03 和 07 是对模型没有影响的特征，可以删除（请注意，在数据设置中，我们选择了 10 个特征，其中只有 7 个与标签/目标具有有意义的关系）。这是 SHAP 的巨大优势，因为没有它，我们就无法深入了解哪些特征实际上正在影响我们的模型。


上面带有下拉菜单的设置允许我们选择要更详细地探索的单个标签。


前 50 个单独示例的力图。




[18]:





print(f"Current Label Shown: {list_of_labels[current_label.value]}\n")

# print the JS visualization code to the notebook
shap.initjs()

shap.force_plot(
    base_value=explainer.expected_value[current_label.value],
    shap_values=shap_values[current_label.value],
    features=X.iloc[0:50, :],
)















Current Label Shown: labels_01




















[18]:










  可视化已省略，未加载 Javascript 库！
你是否在此笔记本中运行了 `initjs()`？如果此笔记本来自其他用户，你还必须信任此笔记本（文件 -> 信任笔记本）。如果你在 github 上查看此笔记本，则 Javascript 已被剥离以确保安全。如果你正在使用 JupyterLab，则此错误是因为尚未编写 JupyterLab 扩展。

 








参考


多输出回归模型格式/构建主要基于 用于多输出回归的深度学习模型