我有一个80%类别变量的机器学习分类问题。如果我想使用一些分类器进行分类,我必须使用一个热编码吗?我可以将数据传递给分类器而不进行编码吗?

我试图做以下的特征选择:

I read the train file: num_rows_to_read = 10000 train_small = pd.read_csv("../../dataset/train.csv", nrows=num_rows_to_read) I change the type of the categorical features to 'category': non_categorial_features = ['orig_destination_distance', 'srch_adults_cnt', 'srch_children_cnt', 'srch_rm_cnt', 'cnt'] for categorical_feature in list(train_small.columns): if categorical_feature not in non_categorial_features: train_small[categorical_feature] = train_small[categorical_feature].astype('category') I use one hot encoding: train_small_with_dummies = pd.get_dummies(train_small, sparse=True)

问题是,第三部分经常卡住,尽管我使用的是一个强大的机器。

因此,如果没有一个热编码,我就无法进行任何特征选择,以确定特征的重要性。

你有什么建议吗?


当前回答

您可以使用numpy。眼睛的功能。

import numpy as np

def one_hot_encode(x, n_classes):
    """
    One hot encode a list of sample labels. Return a one-hot encoded vector for each label.
    : x: List of sample Labels
    : return: Numpy array of one-hot encoded labels
     """
    return np.eye(n_classes)[x]

def main():
    list = [0,1,2,3,4,3,2,1,0]
    n_classes = 5
    one_hot_list = one_hot_encode(list, n_classes)
    print(one_hot_list)

if __name__ == "__main__":
    main()

结果

D:\Desktop>python test.py
[[ 1.  0.  0.  0.  0.]
 [ 0.  1.  0.  0.  0.]
 [ 0.  0.  1.  0.  0.]
 [ 0.  0.  0.  1.  0.]
 [ 0.  0.  0.  0.  1.]
 [ 0.  0.  0.  1.  0.]
 [ 0.  0.  1.  0.  0.]
 [ 0.  1.  0.  0.  0.]
 [ 1.  0.  0.  0.  0.]]

其他回答

Pandas as有内置函数“get_dummies”来获取特定列/秒的热编码。

单热编码的一行代码:

df=pd.concat([df,pd.get_dummies(df['column name'],prefix='column name')],axis=1).drop(['column name'],axis=1)

这对我来说很管用:

pandas.factorize( ['B', 'C', 'D', 'B'] )[0]

输出:

[0, 1, 2, 0]

为了补充其他问题,让我提供如何使用Numpy使用Python 2.0函数:

def one_hot(y_):
    # Function to encode output labels from number indexes 
    # e.g.: [[5], [0], [3]] --> [[0, 0, 0, 0, 0, 1], [1, 0, 0, 0, 0, 0], [0, 0, 0, 1, 0, 0]]

    y_ = y_.reshape(len(y_))
    n_values = np.max(y_) + 1
    return np.eye(n_values)[np.array(y_, dtype=np.int32)]  # Returns FLOATS

行n_values = np.max(y_) + 1可以硬编码,以便在使用小批量的情况下使用足够数量的神经元。

使用此函数的演示项目/教程: https://github.com/guillaume-chevalier/LSTM-Human-Activity-Recognition

您可以使用numpy。眼睛的功能。

import numpy as np

def one_hot_encode(x, n_classes):
    """
    One hot encode a list of sample labels. Return a one-hot encoded vector for each label.
    : x: List of sample Labels
    : return: Numpy array of one-hot encoded labels
     """
    return np.eye(n_classes)[x]

def main():
    list = [0,1,2,3,4,3,2,1,0]
    n_classes = 5
    one_hot_list = one_hot_encode(list, n_classes)
    print(one_hot_list)

if __name__ == "__main__":
    main()

结果

D:\Desktop>python test.py
[[ 1.  0.  0.  0.  0.]
 [ 0.  1.  0.  0.  0.]
 [ 0.  0.  1.  0.  0.]
 [ 0.  0.  0.  1.  0.]
 [ 0.  0.  0.  0.  1.]
 [ 0.  0.  0.  1.  0.]
 [ 0.  0.  1.  0.  0.]
 [ 0.  1.  0.  0.  0.]
 [ 1.  0.  0.  0.  0.]]

我在我的声学模型中使用了这个: 也许这对你的模型有帮助。

def one_hot_encoding(x, n_out):
    x = x.astype(int)  
    shape = x.shape
    x = x.flatten()
    N = len(x)
    x_categ = np.zeros((N,n_out))
    x_categ[np.arange(N), x] = 1
    return x_categ.reshape((shape)+(n_out,))