def pred_result():
vocabulary_size = 400000
time_step = 300
embedding_size = 100
# fit the input information with the trained model
predset = pd.read_csv('check.csv')
texts = []
train = pd.read_csv('train.csv')
texts = predset['text'].astype(str)
text_ml = texts
from DataPrep.Clean_Texts import clean_text
texts = texts.map(lambda x: clean_text(x))
tokenizer_pred = Tokenizer(num_words=vocabulary_size)
tokenizer_pred.fit_on_texts(texts)
encoded_pred = tokenizer_pred.texts_to_sequences(texts)
#print(encoded_docs)
vocab_size_pred = len(tokenizer_pred.word_index) + 1
X_pred = sequence.pad_sequences(encoded_pred,
maxlen=time_step,
padding='post')
#Load models
model1 = load_model('Model_CNN.h5')
model2 = pickle.load(open("NB.sav", 'rb'))
model3 = pickle.load(open("Logistic.sav", 'rb'))
model4 = pickle.load(open("SVM.sav", 'rb'))
y_pred_1 = model1.predict(X_pred)
cv = TfidfVectorizer(max_features=5000)
cleaned = DataClean.cleaned()
X_all = cv.fit_transform(cleaned)
arr_texts = cv.transform(text_ml)
model2 = pickle.load(open("NB.sav", 'rb'))
y_pred_2 = model2.predict(arr_texts)
y_pred_3 = model3.predict(arr_texts)
y_pred_4 = model4.predict(arr_texts)
count = []
#count prediction result and get the reliability index
if math.ceil(y_pred_1[0][0]) == 1:
count.append(1)
if y_pred_2[0] == 1:
count.append(1)
if y_pred_3[0] == 1:
count.append(1)
if y_pred_4[0] == 1:
count.append(1)
return count
from sklearn.model_selection import train_test_split
import numpy as np
import pickle
from sklearn.preprocessing import LabelEncoder
dataVal_Fake_Real = pd.read_csv('fake_or_real_news.csv')
texts = []
texts = dataVal_Fake_Real['text'] #####################################
label = dataVal_Fake_Real['label']
#print(label)
#X=texts.astype(str).values.tolist()
#X=np.reshape(X,(-1,1))
from DataPrep.Clean_Texts import clean_text
X = texts.map(lambda x: clean_text(x))
#print(X)
#label=label.astype(int).values
labelEncoder = LabelEncoder()
encoded_label = labelEncoder.fit_transform(label)
y = np.reshape(encoded_label, (-1, 1))
training_size = int(0.8 * X.shape[0])
X_train = X[:training_size]
y_train = y[:training_size]
X_test = X[training_size:]
y_test = y[training_size:]
#print(X_test)
#print(len(X_train),len(X_test))
import pandas as pd
from DataPrep.Clean_Texts import clean_text
import pickle
dataset = pd.read_csv('balanced_train_data.csv')
#dataset=dataset.head(10)
texts = dataset['comment_text']
#texts=texts.map(lambda x: clean_text(x))
for i in range(len(texts)):
texts[i] = clean_text(texts[i])
label = dataset.iloc[:, 2:8].values
#print(texts[0])
with open('Pickles/pickle_toxic_clean_balanced_train_Xy.pickle', 'wb') as f:
pickle._dump((texts, label), f)
'''pickle_load=open('Pickles/pickle_toxic_clean_balanced_train_Xy.pickle','rb')
X,y=pickle.load(pickle_load)
print(X[0])'''
dataset_test = pd.read_csv('balanced_test_data.csv')
texts_test = dataset_test['comment_text']
#texts=texts.map(lambda x: clean_text(x))
for i in range(len(texts_test)):
texts_test[i] = clean_text(texts_test[i])
label_test = dataset_test.iloc[:, 2:8].values
