ELI5 explain_weights和explain_predictions as pandas DataFrame

Question

如何将ELI5 explain_weights和explain_prediction保存为pandas DataFrame？

下面是玩具代码示例。

# Load the dataset as DataFrame
import pandas as pd
from sklearn.datasets import load_iris
df = pd.DataFrame(load_iris().data, columns=load_iris().feature_names)
df['label'] = load_iris().target

# Divide to X and y and split to train and test sets
X = df.iloc[:,0:4].values
y = df.iloc[:,4].values
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0)

# Define the neural network model
from keras.models import Sequential
from keras.layers import Dense
def baseline_model():
    model = Sequential()
    model.add(Dense(units=8, activation='relu', input_dim=4))
    model.add(Dense(units=3, activation='sigmoid'))
    model.compile(optimizer='adam', loss='categorical_crossentropy', metrics='accuracy'])
    return model

# Wrap the model
from keras.wrappers.scikit_learn import KerasClassifier
my_model = KerasClassifier(build_fn=baseline_model)
my_model.fit(X_train, y_train, batch_size=4, epochs=100, verbose=0)

# Run eli5 explanations
import eli5
from eli5.sklearn import PermutationImportance
perm = PermutationImportance(my_model, random_state=1).fit(X_test, y_test)

# This actually works!
explanation = eli5.formatters.as_dataframe.explain_weights_df(perm, 
feature_names=df.columns[:-1].tolist())

# But this line does not "work", i.e. it produces a NoneType object
explanation_pred = 
eli5.formatters.as_dataframe.explain_prediction_df(estimator=my_model, 
doc=X_test[0])

Answer 1

尝尝这个。这对我很管用。

explanation_pred = eli5.explain_prediction_df(estimator=my_model, doc=X_test.iloc[0])

Answer 2

以防你还在寻找答案。您只需替换以下代码行：

explanation_pred = eli5.formatters.as_dataframe.explain_prediction_df(estimator=my_model, 
doc=X_test[0])

至：

explanation_pred = eli5.explain_prediction_df(estimator=my_model, doc=X_test[0])