我在文本分类问题中有一个数据类型问题

Question

我想为 Kickstarter 活动预测构建深度学习分类器。我的模型部分有问题，但我无法解决。

我的代码：

import pandas as pd
import numpy as np
from sklearn.model_selection import train_test_split
from sklearn.feature_extraction.text import CountVectorizer
from keras.models import Sequential
from keras import layers


df = pd.read_csv('../input/kickstarter-campaigns-dataset/kickstarter_data_full.csv')

df_X = [] # for x class
df_y = [] # for labels

for i in range(len(df)):
    tmp = str(df['blurb'][i]) + " " + str(df['goal'][i]) + " " + str(df['pledged'][i]) + " " + str(df['country'][i]) + " " + str(df['currency'][i]) + " " + str(df['category'][i]) + " " + str(df['spotlight'][i])  
    df_X.append(tmp)
    df_y.append(str(df['SuccessfulBool'][i]))

X_train, X_test, y_train, y_test = train_test_split(df_X, df_y, test_size=0.25, random_state=1000)
vectorizer = CountVectorizer()
vectorizer.fit(X_train)

X_train = vectorizer.transform(X_train)
X_test  = vectorizer.transform(X_test)

input_dim = X_train.shape[1]

model = Sequential()
model.add(layers.Dense(10, input_dim=input_dim, activation='relu'))
model.add(layers.Dense(1, activation='sigmoid'))

model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])
model.summary()

history = model.fit(X_train, y_train,
                     epochs=100,
                     verbose=False,
                     validation_data=(X_test, y_test),
                     batch_size=10)

在这一点上，我得到 ValueError: Failed to find data adapter that can handle input: , ( contains values类型 {""})

我尝试使用 np.asarray 来解决

X_train = np.asarray(X_train)
y_train = np.asarray(y_train)
X_test = np.asarray(X_test)
y_test = np.asarray(y_test)

我得到这个ValueError: Failed to convert a NumPy array to a Tensor (Unsupported object type csr_matrix)。

因此，我使用这个：

np.asarray(X_train).astype(np.float32)
np.asarray(y_train).astype(np.float32)
np.asarray(X_test).astype(np.float32)
np.asarray(y_test).astype(np.float32)

但我得到 ValueError: setting an array element with a sequence。

我试试这个：

X_train = np.expand_dims(X_train, -1)
y_train   = np.expand_dims(y_train, -1)
X_test = np.expand_dims(X_test, -1)
y_test   = np.expand_dims(y_test, -1)

但我在历史部分中不断遇到同样的错误。 ValueError：无法将 NumPy 数组转换为张量（不支持的对象类型 csr_matrix）。

我在 Kaggle 研究 Kickstarter 活动数据集。 https://www.kaggle.com/sripaadsrinivasan/kickstarter-campaigns-dataset

我没有足够的 NLP 信息。我搜索并尝试解决，但我无法解决。这是我的作业。你能帮我解决这个问题吗？

df_X 和 df_y 大小相等，输出如下： x y

Answer 1

您需要在 NN 的顶部添加一个嵌入层来对单词进行矢量化。像这样：

import pandas as pd
import numpy as np
from sklearn.model_selection import train_test_split
from sklearn.feature_extraction.text import CountVectorizer
from keras.preprocessing.text import one_hot
from keras.preprocessing.sequence import pad_sequences
from keras.models import Sequential
from keras import layers


df = pd.read_csv('../input/kickstarter-campaigns-dataset/kickstarter_data_full.csv')

df_X = [] # for x class
df_y = [] # for labels

for i in range(len(df)):
    tmp = str(df['blurb'][i]) + " " + str(df['goal'][i]) + " " + str(df['pledged'][i]) + " " + str(df['country'][i]) + " " + str(df['currency'][i]) + " " + str(df['category'][i]) + " " + str(df['spotlight'][i])  
    df_X.append(tmp)
    df_y.append(str(df['SuccessfulBool'][i]))

vocab_size = 1000
encoded_docs = [one_hot(d, vocab_size) for d in df_X]
max_length = 20
padded_docs = pad_sequences(encoded_docs, maxlen=max_length, padding='post')
X_train, X_test, y_train, y_test = train_test_split(padded_docs, np.array(df_y)[:, None].astype(int), test_size=0.25, random_state=1000)
model = Sequential()
model.add(layers.Embedding(vocab_size, 100, input_length=max_length))
model.add(layers.Flatten())
model.add(layers.Dense(10, activation='relu'))
model.add(layers.Dense(1, activation='sigmoid'))

model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])
print(model.summary())
model.fit(X_train, y_train, 
          epochs=50, 
          verbose=1,
          validation_data=(X_test, y_test),
          batch_size=10)