我如何使用 Keras 对字符串列表进行热编码？

Question

我有一个列表：

code = ['<s>', 'are', 'defined', 'in', 'the', '"editable', 'parameters"', '\n', 'section.', '\n', 'A', 'larger', '`tsteps`', 'value', 'means', 'that', 'the', 'LSTM', 'will', 'need', 'more', 'memory', '\n', 'to', 'figure', 'out']

而且我想转换为一种热编码。我试过了：

to_categorical(code)

我得到一个错误：ValueError: invalid literal for int() with base 10: '<s>'

我做错了什么？

Answer 1

首先尝试将其转换为 numpy 数组：

from numpy import array

然后：

to_categorical(array(code))

Answer 2

keras只支持one-hot-encoding已经integer-encoded的数据。您可以像这样手动 integer-encode 您的字符串：

手动编码

# this integer encoding is purely based on position, you can do this in other ways
integer_mapping = {x: i for i,x in enumerate(code)}

vec = [integer_mapping[word] for word in code]
# vec is
# [0, 1, 2, 3, 16, 5, 6, 22, 8, 22, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25]

使用scikit-learn

from sklearn.preprocessing import LabelEncoder
import numpy as np

code = np.array(code)

label_encoder = LabelEncoder()
vec = label_encoder.fit_transform(code)

# array([ 2,  6,  7,  9, 19,  1, 16,  0, 17,  0,  3, 10,  5, 21, 11, 18, 19,
#         4, 22, 14, 13, 12,  0, 20,  8, 15])

您现在可以将其输入 keras.utils.to_categorical:

from keras.utils import to_categorical

to_categorical(vec)

Answer 3

改为使用

pandas.get_dummies(y_train)

Answer 4

tf.keras.layers.CategoryEncoding

在 TF 2.6.0 中，可以使用 tf.keras.layers.CategoryEncoding , tf.keras.layers.StringLookup, and tf.keras.layers.IntegerLookup.

实现 One Hot Encoding (OHE) 或 Multi Hot Encoding (MHE)

我认为这种方式在TF 2中是不合理的。4.x所以它一定是在之后实现的。

请参阅 Classify structured data using Keras preprocessing layers 以了解实际实施。

def get_category_encoding_layer(name, dataset, dtype, max_tokens=None):
  # Create a layer that turns strings into integer indices.
  if dtype == 'string':
    index = layers.StringLookup(max_tokens=max_tokens)
  # Otherwise, create a layer that turns integer values into integer indices.
  else:
    index = layers.IntegerLookup(max_tokens=max_tokens)

  # Prepare a `tf.data.Dataset` that only yields the feature.
  feature_ds = dataset.map(lambda x, y: x[name])

  # Learn the set of possible values and assign them a fixed integer index.
  index.adapt(feature_ds)

  # Encode the integer indices.
  encoder = layers.CategoryEncoding(num_tokens=index.vocabulary_size())

  # Apply multi-hot encoding to the indices. The lambda function captures the
  # layer, so you can use them, or include them in the Keras Functional model later.
  return lambda feature: encoder(index(feature))

我如何使用 Keras 对字符串列表进行热编码？

How can I one hot encode a list of strings with Keras?

python

keras

one-hot-encoding

手动编码

使用scikit-learn

tf.keras.layers.CategoryEncoding