for n in [2000, 5000, 10000, 15000, 20000]:
model = ClassifierModel("Count Max Feature {}".format(n),
CountVectorizer(max_features=n))
models.append(model)
for n in [2000, 5000, 10000, 15000, 20000]:
model = ClassifierModel("Count Max Feature {}".format(n),
TfidfVectorizer(max_features=n))
models.append(model)
for n in [500, 700, 800, 900, 1000, 2000, 3000, 4000, 5000]:
for ngram in [('Bigram', (1, 2)), ("Trigram", (1, 3))]:
model = ClassifierModel(
"Count {0} + Max Feature {1}".format(ngram[0], n),
CountVectorizer(ngram_range=ngram[1], max_features=n))
models.append(model)
for n in [500, 700, 800, 900, 1000, 2000, 3000, 4000, 5000]:
for ngram in [('Bigram', (1, 2)), ("Trigram", (1, 3))]:
model = ClassifierModel(
"Count {0} + Max Feature {1}".format(ngram[0], n),
TfidfVectorizer(ngram_range=ngram[1], max_features=n))
models.append(model)
for model in models:
model.load_data(X_train, y_train, X_test, y_test)
model.fit_transform()
model.train(clf_model=MultinomialNB())
model.evaluate(model_name="MultinomialNB")
models.append(model)
for n in [2000, 5000, 10000, 15000, 20000]:
model = ClassifierModel(
"Tfidf Max Feature {}".format(n),
TfidfVectorizer(max_features=n)
)
models.append(model)
for n in [500, 700, 800, 900, 1000, 2000, 3000, 4000, 5000]:
for ngram in [('Bigram', (1, 2)), ("Trigram", (1, 3))]:
model = ClassifierModel(
"Count {0} + Max Feature {1}".format(ngram[0], n),
CountVectorizer(ngram_range=ngram[1], max_features=n)
)
models.append(model)
for n in [500, 700, 800, 900, 1000, 2000, 3000, 4000, 5000]:
for ngram in [('Bigram', (1, 2)), ("Trigram", (1, 3))]:
model = ClassifierModel(
"Tfidf {0} + Max Feature {1}".format(ngram[0], n),
TfidfVectorizer(ngram_range=ngram[1], max_features=n)
)
models.append(model)
for model in models:
model.load_data(X_train, y_train, X_test, y_test)
model.fit_transform()
model.train(clf_model=LinearSVC())
model.evaluate(model_name="LinearSVC")
for n in [2000, 5000, 10000, 15000, 20000]:
model = ClassifierModel("Count Max Feature {}".format(n),
CountVectorizer(max_features=n))
models.append(model)
for n in [2000, 5000, 10000, 15000, 20000]:
model = ClassifierModel("Tfidf Max Feature {}".format(n),
TfidfVectorizer(max_features=n))
models.append(model)
for n in [500, 700, 800, 900, 1000, 2000, 3000, 4000, 5000]:
for ngram in [('Bigram', (1, 2)), ("Trigram", (1, 3))]:
model = ClassifierModel(
"Count {0} + Max Feature {1}".format(ngram[0], n),
CountVectorizer(ngram_range=ngram[1], max_features=n))
models.append(model)
for n in [500, 700, 800, 900, 1000, 2000, 3000, 4000, 5000]:
for ngram in [('Bigram', (1, 2)), ("Trigram", (1, 3))]:
model = ClassifierModel(
"Tfidf {0} + Max Feature {1}".format(ngram[0], n),
TfidfVectorizer(ngram_range=ngram[1], max_features=n))
models.append(model)
for model in models:
model.load_data(X_train, y_train, X_test, y_test)
model.fit_transform()
model.train(clf_model=SVC())
model.evaluate(model_name="SVC")
models.append(model)
for n in [2000, 5000, 10000, 15000, 20000]:
model = ClassifierModel(
"Tfidf Max Feature {}".format(n),
TfidfVectorizer(max_features=n)
)
models.append(model)
for n in [500, 700, 800, 900, 1000, 2000, 3000, 4000, 5000]:
for ngram in [('Bigram', (1, 2)), ("Trigram", (1, 3))]:
model = ClassifierModel(
"Count {0} + Max Feature {1}".format(ngram[0], n),
CountVectorizer(ngram_range=ngram[1], max_features=n)
)
models.append(model)
for n in [500, 700, 800, 900, 1000, 2000, 3000, 4000, 5000]:
for ngram in [('Bigram', (1, 2)), ("Trigram", (1, 3))]:
model = ClassifierModel(
"Tfidf {0} + Max Feature {1}".format(ngram[0], n),
TfidfVectorizer(ngram_range=ngram[1], max_features=n)
)
models.append(model)
for model in models:
model.load_data(X_train, y_train, X_test, y_test)
model.fit_transform()
model.train(clf_model=LogisticRegression())
model.evaluate(model_name="LogisticRegression")