计算互信息(MI)值时遇ValueError：样本数为0的解决求助

在计算互信息值时反复触发以下错误：

ValueError: Found array with 0 sample(s) (shape=(0, 1)) while a minimum of 1 is required

我的代码： #

X_new = X.copy()
X_new = X_new.fillna(0)
y = data.SalePrice


def make_mi_scores(X, y):
    X = X.copy()
    for colname in X.select_dtypes(["object", "category"]):
        X[colname], _ = X[colname].factorize()       
    discrete_features = [pd.api.types.is_integer_dtype(t) for t in X.dtypes]   
    mi_scores = mutual_info_regression(X, y, discrete_features=discrete_features, random_state=0)
    mi_scores = pd.Series(mi_scores, name="MI Scores", index=X.columns)
    mi_scores = mi_scores.sort_values(ascending=False)
    return mi_scores

def plot_mi_scores(scores):
    scores = scores.sort_values(ascending=True)
    width = np.arange(len(scores))
    ticks = list(scores.index)
    plt.barh(width, scores)
    plt.yticks(width, ticks)
    plt.title("Mutual Information Scores")
    
plt.figure(dpi=100, figsize=(8, 5))
plot_mi_scores(make_mi_scores(X_new,y))

报错堆栈： #

ValueError                                Traceback (most recent call last)
/tmp/ipykernel_19/1575243112.py in <module>
     42 
     43 plt.figure(dpi=100, figsize=(8, 5))
---> 44 plot_mi_scores(make_mi_scores(X_new,y))

/tmp/ipykernel_19/1575243112.py in make_mi_scores(X, y)
     28     print(X.isnull().any().any())
     29     print(y.isnull().any().any())
---> 30     mi_scores = mutual_info_regression(X, y, discrete_features=discrete_features, random_state=0)
     31     mi_scores = pd.Series(mi_scores, name="MI Scores", index=X.columns)
     32     mi_scores = mi_scores.sort_values(ascending=False)

/opt/conda/lib/python3.7/site-packages/sklearn/feature_selection/_mutual_info.py in mutual_info_regression(X, y, discrete_features, n_neighbors, copy, random_state)
    382            of a Random Vector", Probl. Peredachi Inf., 23:2 (1987), 9-16
    383     """
---> 384     return _estimate_mi(X, y, discrete_features, False, n_neighbors, copy, random_state)
    385 
    386 

/opt/conda/lib/python3.7/site-packages/sklearn/feature_selection/_mutual_info.py in _estimate_mi(X, y, discrete_features, discrete_target, n_neighbors, copy, random_state)
    300     mi = [
    301         _compute_mi(x, y, discrete_feature, discrete_target, n_neighbors)
---> 302         for x, discrete_feature in zip(_iterate_columns(X), discrete_mask)
    303     ]
    304 

/opt/conda/lib/python3.7/site-packages/sklearn/feature_selection/_mutual_info.py in <listcomp>(.0)
    300     mi = [
    301         _compute_mi(x, y, discrete_feature, discrete_target, n_neighbors)
---> 302         for x, discrete_feature in zip(_iterate_columns(X), discrete_mask)
    303     ]
    304 

/opt/conda/lib/python3.7/site-packages/sklearn/feature_selection/_mutual_info.py in _compute_mi(x, y, x_discrete, y_discrete, n_neighbors)
    160         return mutual_info_score(x, y)
    161     elif x_discrete and not y_discrete:
---> 162         return _compute_mi_cd(y, x, n_neighbors)
    163     elif not x_discrete and y_discrete:
    164         return _compute_mi_cd(x, y, n_neighbors)

/opt/conda/lib/python3.7/site-packages/sklearn/feature_selection/_mutual_info.py in _compute_mi_cd(c, d, n_neighbors)
    137     radius = radius[mask]
    138 
---> 139     kd = KDTree(c)
    140     m_all = kd.query_radius(c, radius, count_only=True, return_distance=False)
    141     m_all = np.array(m_all) - 1.0

sklearn/neighbors/_binary_tree.pxi in sklearn.neighbors._kd_tree.BinaryTree.__init__()

/opt/conda/lib/python3.7/site-packages/sklearn/utils/validation.py in check_array(array, accept_sparse, accept_large_sparse, dtype, order, copy, force_all_finite, ensure_2d, allow_nd, ensure_min_samples, ensure_min_features, estimator)
    806                 "Found array with %d sample(s) (shape=%s) while a"
    807                 " minimum of %d is required%s."
---> 808                 % (n_samples, array.shape, ensure_min_samples, context)
    809             )
    810 

ValueError: Found array with 0 sample(s) (shape=(0, 1)) while a minimum of 1 is required.

解决方法 #

1. 过滤无信息量特征（核心修复）

错误根源是数据中存在全部值相同的离散特征，这类特征在计算互信息时会导致KDTree构建失败。在make_mi_scores函数开头添加过滤逻辑：

def make_mi_scores(X, y):
    X = X.copy()
    # 过滤掉所有值完全相同的特征
    X = X.loc[:, X.nunique() > 1]
    for colname in X.select_dtypes(["object", "category"]):
        X[colname], _ = X[colname].factorize()       
    discrete_features = [pd.api.types.is_integer_dtype(t) for t in X.dtypes]   
    mi_scores = mutual_info_regression(X, y, discrete_features=discrete_features, random_state=0)
    mi_scores = pd.Series(mi_scores, name="MI Scores", index=X.columns)
    mi_scores = mi_scores.sort_values(ascending=False)
    return mi_scores

2. 验证输入数据完整性

确保特征集和目标变量没有空样本：

# 在调用函数前添加检查
assert len(X_new) > 0, "特征集样本数为0"
assert len(y) > 0, "目标变量样本数为0"

3. 优化离散特征判断逻辑

is_integer_dtype可能误判用整数存储的连续值，可改用sklearn自动判断离散特征：

# 修改mutual_info_regression的参数
mi_scores = mutual_info_regression(X, y, discrete_features='auto', random_state=0)

内容的提问来源于stack exchange，提问作者Tottle