在scikit-learn中使用预先训练好的单词嵌入

from sklearn.feature_extraction.text import CountVectorizer
from sklearn.feature_extraction.text import TfidfTransformer
from sklearn.pipeline import Pipeline
from sklearn.svm import LinearSVC

X_train = ['I love this movie', 'This movie is great', 'I dislike this movie']
y_train = ['positive', 'positive', 'negative']

pipeline = Pipeline([
    ('vect', CountVectorizer()),
    ('tfidf', TfidfTransformer()),
    ('clf', LinearSVC())
])

# 使用预先训练好的单词嵌入文件（例如GloVe）
word_embeddings_file = 'path/to/word_embeddings.txt'

# 设置CountVectorizer的vocabulary参数为预先训练好的单词嵌入
with open(word_embeddings_file, 'r', encoding='utf-8') as f:
    word_embeddings = {}
    for line in f:
        values = line.split()
        word = values[0]
        embedding = np.asarray(values[1:], dtype='float32')
        word_embeddings[word] = embedding

pipeline.named_steps['vect'].vocabulary_ = word_embeddings

pipeline.fit(X_train, y_train)
y_pred = pipeline.predict(X_test)