麻烦实施伯努利朴素贝叶斯分类器

Question

我正在尝试从Bernoulli Naive Bayes库实现scikit-learn分类器进行文本分类。但我坚持这个错误

ValueError：预期的2D数组，改为获得1D数组：

如果数据具有单个要素，则使用array.reshape（-1,1）重新整形数据;如果数据包含单个样本，则使用array.reshape（1，-1）重新整形数据。

详细错误

Traceback (most recent call last):
  File "BNB.py", line 27, in <module>
    clf.fit(train_data, train_labels)
  File "/home/atinesh/.local/lib/python3.6/site-packages/sklearn/naive_bayes.py", line 579, in fit
    X, y = check_X_y(X, y, 'csr')
  File "/home/atinesh/.local/lib/python3.6/site-packages/sklearn/utils/validation.py", line 573, in check_X_y
    ensure_min_features, warn_on_dtype, estimator)
  File "/home/atinesh/.local/lib/python3.6/site-packages/sklearn/utils/validation.py", line 441, in check_array
    "if it contains a single sample.".format(array))
ValueError: Expected 2D array, got 1D array instead:
array=['Apple' 'Banana' 'Cherry' 'Grape' 'Guava' 'Lemon' 'Mangos' 'Orange'
 'Strawberry' 'Watermelon' 'Potato' 'Spinach' 'Carrot' 'Onion' 'Cabbage'
 'Barccoli' 'Tomatoe' 'Pea' 'Cucumber' 'Eggplant'].
Reshape your data either using array.reshape(-1, 1) if your data has a single feature or array.reshape(1, -1) if it contains a single sample.

"B NB.朋友"

from sklearn.naive_bayes import BernoulliNB

dataPos = ['Apple', 'Banana', 'Cherry', 'Grape', 'Guava', 'Lemon', 'Mangos',
            'Orange', 'Strawberry', 'Watermelon']

dataNeg = ['Potato', 'Spinach', 'Carrot', 'Onion', 'Cabbage', 'Barccoli', 
            'Tomatoe', 'Pea', 'Cucumber', 'Eggplant']

def get_data():
    examples = []
    labels   = []

    for item in dataPos:
        examples.append(item)
        labels.append('positive')

    for item in dataNeg:
        examples.append(item)
        labels.append('negative')

    return examples, labels

train_data, train_labels = get_data()

# Train
clf = BernoulliNB()
clf.fit(train_data, train_labels)

# Predict
print(clf.predict('Apple Banana'))
print(clf.predict_proba('Apple Banana'))

Answer 1

我建议在sklearn中使用LabelBinarizer

from sklearn.naive_bayes import BernoulliNB
import numpy as np
from sklearn import preprocessing

dataPos = ['Apple', 'Banana', 'Cherry', 'Grape', 'Guava', 'Lemon', 'Mangos',
                       'Orange', 'Strawberry', 'Watermelon']

dataNeg = ['Potato', 'Spinach', 'Carrot', 'Onion', 'Cabbage', 'Barccoli',
                       'Tomatoe', 'Pea', 'Cucumber', 'Eggplant']

Y=[0]*10+[1]*10
Y=np.array(Y)

lb = preprocessing.LabelBinarizer()
X = lb.fit_transform(dataPos+dataNeg)
clf = BernoulliNB()
clf.fit(X, Y)

test_sample = lb.transform([['Apple'],['Banana'],['Spinach']])
print clf.predict(test_sample)

您的代码错误，因为在执行clf.fit(X,Y)时，X需要是2d数组。每行对应一个特征向量。

Answer 2

如果将简单的python列表传递给scikit_learn，它将被解释为shape（n，）数组。您可能想要做的是将示例和标签的列表转换为numpy数组，并将它们重新整形/调整为形状为（n，1）的线矢量。例如：

import numpy as np

examples = np.array(['Apple', 'Banana', 'Cherry', 'Grape', 'Guava', 'Lemon', 'Mangos','Orange', 'Strawberry', 'Watermelon'])
examples.shape  # returns (10, ), a 1D-array
examples.resize((10,1))
examples.shape  # returns (10, 1), which is a 2-D array

或者对于更简单的解决方案，您可以简单地提供fit方法：

clf.fit([train_data], [train_labels])

但是，由于您已经有一个专门的方法来格式化数据，为什么不在那里使用numpy并返回具有正确尺寸的列表。

希望这有助于您的努力。