SciKit TypeError：根据规则'safe'，无法将数组数据从dtype（'float64'）转换为dtype（'u32'）

Question

我正在使用带有以下代码的sklearn库，并获得标题中提到的错误。如果我使用非线性分类器（例如DecisionTreeClassifier或KNeighborsClassifier）但使用像rayD和PassiveAggressive这样的linear_model分类器，则代码可以正常运行。我已经尝试将数组内容映射到浮点数，但它不起作用。数组的所有内容都是'numpy.unicode_'。我很感激帮助解决这个问题。

代码：

from io import open
from os import listdir
import numpy as np

path = "address"
months = listdir(path)
for month in months:
    if "_sanitized" in month:
        files = listdir (path + month)
        for file in files:
            temp_path = path + month + "\\" + file
            text = open (temp_path, "r").read()
            lines = text.split("\n")
            attr_train = list()
            label_train = list()
            go = False
            for line in lines:
                if line != "" and not go:
                    row = line.split()
                    label_train.append(row[-1])
                    go = (row[-1] != u'-1')
                    del row[-1]
                    attr_train.append(row)
            label_train_arr = np.asarray(label_train)
        attr_train_arr = np.asarray(attr_train)
        break
    break

from sklearn.model_selection import train_test_split
from sklearn import preprocessing

le = preprocessing.LabelEncoder()

for arr in range(len(attr_train_arr)):
    newArr = le.fit_transform(attr_train_arr[arr])
    attr_train_arr[arr] = newArr

label_train_arr = map(float, label_train_arr)
for arr in attr_train_arr:
    arr = map(float, arr)
x = attr_train_arr
y = label_train_arr

X_train, X_test, Y_train, Y_test = train_test_split(x, y, test_size = .5)
print "after split"

from sklearn.linear_model import SGDClassifier
#my_classifier = SGDClassifier()

from sklearn.linear_model import PassiveAggressiveClassifier
my_classifier = PassiveAggressiveClassifier()

my_classifier.fit(X_train, Y_train)
print "after fit"

predictions = my_classifier.predict(X_test)