def get_feature_importance(language: str, embeddings_path: str,
platform_folder: str, n_estimators: int,
platform: str, dataset: str):
print("#################")
print(dataset)
print()
t = Model(language=language,
embeddings_path=embeddings_path,
platform_folder=platform_folder)
feature_importances = t.feature_importance(n_estimators=n_estimators,
platform=platform,
dataset=dataset)
print(feature_importances)
print()
from sklearn.svm import SVC
from sklearn.neighbors import KNeighborsClassifier
from sklearn.naive_bayes import GaussianNB, MultinomialNB
from sklearn.ensemble import RandomForestClassifier
from sklearn.neighbors import KNeighborsClassifier
from fakenews_detector.model import Model
dataset = "tweets_br"
platform = 'Twitter'
language = "pt"
tree_estimators = np.arange(1, 1001, 50)
t = Model(language='pt',
embeddings_path='embeddings/pt/model.txt',
platform_folder='datasets/Twitter/tweets_br/')
##########################################################################################################
print("TESTING WITH LTO")
print()
print()
# FEATURES
print("FEATURES + GAUSSIANNB")
t.custom_gridsearch(dataset=dataset,
platform=platform,
parameters={},
language=language,
clf="gnb",
feature_set=Model.FEATURES_CODE)
from sklearn.svm import SVC
from sklearn.neighbors import KNeighborsClassifier
from sklearn.naive_bayes import GaussianNB, MultinomialNB
from sklearn.ensemble import RandomForestClassifier
from sklearn.neighbors import KNeighborsClassifier
from fakenews_detector.model import Model
language = 'bg'
platform = 'Websites'
dataset = 'btv-lifestyle'
tree_estimators = np.arange(1, 1001, 50)
t = Model(language='bg',
embeddings_path='embeddings/bg/model.txt',
platform_folder='datasets/Websites/btv-lifestyle/')
# FEATURES
print("FEATURES + GAUSSIANNB")
t.classify(dataset=dataset,
platform=platform,
language=language,
clf=GaussianNB(),
parameters={},
feature_set=Model.FEATURES_CODE,
kfold=5)
print("------------------------------------------------------")
print("FEATURES + KNN")
t.classify(dataset=dataset,
from sklearn.svm import SVC
from sklearn.neighbors import KNeighborsClassifier
from sklearn.naive_bayes import GaussianNB, MultinomialNB
from sklearn.ensemble import RandomForestClassifier
from sklearn.neighbors import KNeighborsClassifier
from fakenews_detector.model import Model
dataset = "tweets_br"
platform = 'Twitter'
language = "pt"
tree_estimators = np.arange(1, 1001, 50)
t = Model(language='pt',
embeddings_path='embeddings/pt/model.txt',
platform_folder='datasets/Twitter/tweets_br/')
##########################################################################################################
print("TESTING WITH LTO")
print()
print()
# FEATURES
print("BOW + SVC")
fscore_topic_svm, _, _, _, _ = t.classify_lto('Twitter', 'tweets_br', 'pt',
SVC(kernel='poly'),
Model.BOW_CODE)
print("------------------------------------------------------")