有没有办法将列表用作 DataFrame 中的值？

Question

我正在处理著名的 Kaggle 挑战 "House prices"。我想用 sklearn.linear_model LinearRegression

训练我的数据集

阅读以下文章后： https://developers.google.com/machine-learning/crash-course/representation/feature-engineering

我编写了一个函数，将我的火车 DataFrame 中的所有字符串值转换为列表。例如，原始特征值可能看起来像这样 [Ex, Gd, Ta, Po]，转换后它会像这样：[1,0,0,0] [0,1,0,0] [0, 0,1,0] [0,0,0,1].

当我尝试训练我的数据时，出现以下错误：

Traceback (most recent call last): File "C:/Users/Owner/PycharmProjects/HousePrices/main.py", line 27, in linereg.fit(train_df, target) File "C:\Users\Owner\PycharmProjects\HousePrices\venv\lib\site-packages\sklearn\linear_model\base.py", line 458, in fit y_numeric=True, multi_output=True) File "C:\Users\Owner\PycharmProjects\HousePrices\venv\lib\site-packages\sklearn\utils\validation.py", line 756, in check_X_y estimator=estimator) File "C:\Users\Owner\PycharmProjects\HousePrices\venv\lib\site-packages\sklearn\utils\validation.py", line 567, in check_array array = array.astype(np.float64) ValueError: setting an array element with a sequence.

只有当我按照我的解释转换某些列时才会发生这种情况。

有什么方法可以训练以向量为值的线性回归模型吗？

这是我的转换函数：

def feature_to_boolean_vector(df, feature_name, new_name):
    vectors_list = [] #each tuple will represent an option
    feature_options = df[feature_name].unique()
    feature_options_length = len(feature_options)

    # creating a list the size of feature_options_length, all 0's
    list_to_be_vector = [0 for i in range(feature_options_length)]

    for i in range(feature_options_length):
        list_to_be_vector[i] = 1 # inserting 1 representing option number i
        vectors_list.append(list_to_be_vector.copy())
        list_to_be_vector[i] = 0

    mapping = dict(zip(feature_options, vectors_list)) # dict from values to vectors
    df[new_name] = df[feature_name].map(mapping)
    df.drop([feature_name], axis=1, inplace=True)

这是我的火车尝试（预处理后）：

linereg = LinearRegression()
linereg.fit(train_df, target)

提前致谢。

Answer 1

LinearRegression 不支持列表功能。我看到你在使用 one-hot，你可以将每个维度用作一列特征。相比之下，您可以使用更简单的方法 pd.get_dummies in pandas.

print(df['feature'])
0    Ex
1    Gd
2    Ta
3    Po
Name: feature, dtype: object

df = pd.get_dummies(df['feature'])
print(df)
   Ex  Gd  Po  Ta
0   1   0   0   0
1   0   1   0   0
2   0   0   0   1
3   0   0   1   0

有没有办法将列表用作 DataFrame 中的值？

Is there a way to use Lists as values in a DataFrame?

python

linear-regression

sklearn-pandas