logging.getLogger("pyhive").setLevel(logging.CRITICAL) # avoid excessive logs
logger.setLevel(logging.INFO)
if __name__ == "__main__":
data = m.Posts()
scorer = make_scorer(fbeta_score, beta=2)
trans_os = {'translate': [0.9], 'oversample': [0.9]}
TARGET_LABELS = [
'label_discriminating', 'label_inappropriate',
'label_sentimentnegative', 'label_needsmoderation'
]
embedding_dict_glove = transformers.load_embedding_vectors(
embedding_style='glove', file="./embeddings/glove_vectors.txt")
embedding_dict_word2vec = transformers.load_embedding_vectors(
embedding_style='word2vec', file="./embeddings/word2vec_vectors.txt")
preps = {
'norm':
lambda x: cleaning.series_apply_chaining(x, [cleaning.normalize]),
'glove':
transformers.MeanEmbeddingVectorizer(
embedding_dict=embedding_dict_glove).transform,
'word2vec':
transformers.MeanEmbeddingVectorizer(
embedding_dict=embedding_dict_word2vec).transform,
}
vecs = {
'count': CountVectorizer(),
'translate': [0.8, 0.9, 1.0],
'oversample': [0.8, 0.9, 1.0]
}
TARGET_LABELS = [
'label_argumentsused',
'label_discriminating',
'label_inappropriate',
'label_offtopic',
'label_personalstories',
'label_possiblyfeedback',
'label_sentimentnegative',
'label_sentimentpositive',
]
embedding_dict_glove = transformers.load_embedding_vectors(
embedding_style='glove')
embedding_dict_word2vec = transformers.load_embedding_vectors(
embedding_style='word2vec')
preps = {
'norm':
lambda x: cleaning.series_apply_chaining(x, [cleaning.normalize]),
'stem':
lambda x: cleaning.series_apply_chaining(
x, [cleaning.normalize, cleaning.stem_germ]),
'lem':
lambda x: cleaning.series_apply_chaining(
x, [cleaning.normalize, cleaning.lem_germ]),
'glove':
transformers.MeanEmbeddingVectorizer(
embedding_dict=embedding_dict_glove).transform,