def run(dataset, hyperparameters, metrics, fname=None):
# # Load Resources
word2vec = None
if hyperparameters['model'] != 'rand':
word2vec = load_word2vec()
# # Load Dataset
df = load_dataset(dataset[0], **dataset[1])
# # Preprocess
df['clean_tweets'] = df.tweet.apply(
TweetPreprocessor(normalize=['link', 'mention']).preprocess)
df['tokens'] = df.clean_tweets.apply(TweetTokenizer().tokenize)
X_train, X_dev, X_test, y_train, y_dev, y_test = train_dev_test_split(
df.tokens, df.label)
# # Train
clf = NeuralNetClassifier(module=TextCNN,
corpus=df.tokens,
word_vectors=word2vec,
metrics=metrics,
**hyperparameters)
clf.fit(X_train, y_train, validation_data=(X_dev, y_dev))
# # Predict
y_pred = clf.predict(X_test)
# # Evaluate
pprint(
dict(dataset=dataset,
hyperparameters=hyperparameters,
scores={
scorer: get_score_func(scorer)(y_test, y_pred)
for scorer in metrics
}))
# # Save to file
X_test['pred'] = y_pred
X_test.to_excel(scratch_path('predictions_%s.xlsx' % fname))
def run_all(**kwargs):
word2vec = load_word2vec()
print("Individual Feature Comparison")
for dataset in kwargs['DATASETS']:
for features in kwargs['FEATURES']:
run(
dataset=dataset,
features=features,
metrics=['accuracy', 'precision', 'recall', 'f1', 'roc_auc'],
word2vec=word2vec,
)
import numpy as np
import pandas as pd
from sklearn.cross_decomposition import CCA
from tklearn.feature_extraction.hatespeech import load_hatebase, HatebaseVectorizer
from tklearn.text.word_vec import load_word2vec
hatebase = load_hatebase()
word2vec = load_word2vec()
embeddings = {}
for w in hatebase.keys():
if w in word2vec.vocab:
embeddings[w] = [word2vec.word_vec(w)]
hv = HatebaseVectorizer(features=['average_offensiveness'])
hv.fit(None)
C1 = np.array([v[0] for w, v in embeddings.items()])
C1.shape
C2 = hv.feature_vectors.loc[[hv.index[w] for w in embeddings.keys()]]
C2.shape
cca = CCA(n_components=25)
cca.fit(C2, C1)
embeddings = {}