fakenews_detection.py
# -*- coding: utf-8 -*-
"""fakenews_seq_classification.ipynb
Automatically generated by Colaboratory.
Original file is located at
https://colab.research.google.com/drive/1e_3Zn4mPSYaMvRvLeOtA8AYXqOSbgkgc
"""
!pip install -q kaggle
from google.colab import files
files.upload()
!rm -rf ~/.kaggle
!mkdir ~/.kaggle
!cp kaggle.json ~/.kaggle/
!chmod 600 ~/.kaggle/kaggle.json
!kaggle competitions download -c fake-news
!unzip test.csv.zip
!unzip train.csv.zip
!pip install gdown
# download from Google Drive
!gdown "https://drive.google.com/uc?id=178f_VkNxccNidap-5-uffXUW475pAuPy&confirm=t"
!unzip fake-news.zip
### Import all library
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
import nltk
nltk.download('stopwords')
nltk.download('wordnet')
# load the dataset
news_d = pd.read_csv("train.csv")
submit_test = pd.read_csv("test.csv")
## Shape and colums of train dataset
print(" Shape of News data :: ", news_d.shape)
print(" News data columns", news_d.columns)
## by using df.head(),We can immediately familiarize ourselves with the dataset.
news_d.head()
#Text Word startistics: min.mean, max and interquartile range
txt_length = news_d.text.str.split().str.len()
txt_length.describe()
#Title statistics
title_length = news_d.title.str.split().str.len()
title_length.describe()
sns.countplot(x="label", data=news_d);
print("1: Unreliable")
print("0: Reliable")
print("Distribution of labels:")
print(news_d.label.value_counts());
print(round(news_d.label.value_counts(normalize=True),2)*100);
# Constants that are used to sanitize the datasets
column_n = ['id', 'title', 'author', 'text', 'label']
remove_c = ['id','author']
categorical_features = []
target_col = ['label']
text_f = ['title', 'text']
# Clean Datasets
import nltk
from nltk.corpus import stopwords
import re
from nltk.stem.porter import PorterStemmer
from collections import Counter
ps = PorterStemmer()
wnl = nltk.stem.WordNetLemmatizer()
stop_words = stopwords.words('english')
stopwords_dict = Counter(stop_words)
# Removed unused clumns
def remove_unused_c(df,column_n=remove_c):
df = df.drop(column_n,axis=1)
return df
# Impute null values with None
def null_process(feature_df):
for col in text_f:
feature_df.loc[feature_df[col].isnull(), col] = "None"
return feature_df
def clean_dataset(df):
# remove unused column
df = remove_unused_c(df)
#impute null values
df = null_process(df)
return df
# Cleaning text from unused characters
def clean_text(text):
text = str(text).replace(r'http[\w:/\.]+', ' ') # removing urls
text = str(text).replace(r'[^\.\w\s]', ' ') # remove everything but characters and punctuation
text = str(text).replace('[^a-zA-Z]', ' ')
text = str(text).replace(r'\s\s+', ' ')
text = text.lower().strip()
#text = ' '.join(text)
return text
## Nltk Preprocessing include:
# Stop words, Stemming and Lemmetization
# For our project we use only Stop word removal
def nltk_preprocess(text):
text = clean_text(text)
wordlist = re.sub(r'[^\w\s]', '', text).split()
#text = ' '.join([word for word in wordlist if word not in stopwords_dict])
#text = [ps.stem(word) for word in wordlist if not word in stopwords_dict]
text = ' '.join([wnl.lemmatize(word) for word in wordlist if word not in stopwords_dict])
return text
# Perform data cleaning on train and test dataset by calling clean_dataset function
df = clean_dataset(news_d)
# apply preprocessing on text through apply method by calling the function nltk_preprocess
df["text"] = df.text.apply(nltk_preprocess)
# apply preprocessing on title through apply method by calling the function nltk_preprocess
df["title"] = df.title.apply(nltk_preprocess)
# Dataset after cleaning and preprocessing step
df.head()
from wordcloud import WordCloud, STOPWORDS
import matplotlib.pyplot as plt
# initialize the word cloud
wordcloud = WordCloud( background_color='black', width=800, height=600)
# generate the word cloud by passing the corpus
text_cloud = wordcloud.generate(' '.join(df['text']))
# plotting the word cloud
plt.figure(figsize=(20,30))
plt.imshow(text_cloud)
plt.axis('off')
plt.show()
true_n = ' '.join(df[df['label']==0]['text'])
wc = wordcloud.generate(true_n)
plt.figure(figsize=(20,30))
plt.imshow(wc)
plt.axis('off')
plt.show()
fake_n = ' '.join(df[df['label']==1]['text'])
wc= wordcloud.generate(fake_n)
plt.figure(figsize=(20,30))
plt.imshow(wc)
plt.axis('off')
plt.show()
def plot_top_ngrams(corpus, title, ylabel, xlabel="Number of Occurences", n=2):
"""Utility function to plot top n-grams"""
true_b = (pd.Series(nltk.ngrams(corpus.split(), n)).value_counts())[:20]
true_b.sort_values().plot.barh(color='blue', width=.9, figsize=(12, 8))
plt.title(title)
plt.ylabel(ylabel)
plt.xlabel(xlabel)
plt.show()
plot_top_ngrams(true_n, 'Top 20 Frequently Occuring True news Bigrams', "Bigram", n=2)
plot_top_ngrams(fake_n, 'Top 20 Frequently Occuring Fake news Bigrams', "Bigram", n=2)
plot_top_ngrams(true_n, 'Top 20 Frequently Occuring True news Trigrams', "Trigrams", n=3)
plot_top_ngrams(fake_n, 'Top 20 Frequently Occuring Fake news Trigrams', "Trigrams", n=3)
"""# Fine-tuning BERT"""
!pip install transformers
import torch
from transformers.file_utils import is_tf_available, is_torch_available, is_torch_tpu_available
from transformers import BertTokenizerFast, BertForSequenceClassification
from transformers import Trainer, TrainingArguments
import numpy as np
from sklearn.model_selection import train_test_split
import random
def set_seed(seed: int):
"""
Helper function for reproducible behavior to set the seed in ``random``, ``numpy``, ``torch`` and/or ``tf`` (if
installed).
Args:
seed (:obj:`int`): The seed to set.
"""
random.seed(seed)
np.random.seed(seed)
if is_torch_available():
torch.manual_seed(seed)
torch.cuda.manual_seed_all(seed)
# ^^ safe to call this function even if cuda is not available
if is_tf_available():
import tensorflow as tf
tf.random.set_seed(seed)
set_seed(1)
# the model we gonna train, base uncased BERT
# check text classification models here: https://huggingface.co/models?filter=text-classification
model_name = "bert-base-uncased"
# max sequence length for each document/sentence sample
max_length = 512
# load the tokenizer
tokenizer = BertTokenizerFast.from_pretrained(model_name, do_lower_case=True)
news_df = news_d[news_d['text'].notna()]
news_df = news_df[news_df["author"].notna()]
news_df = news_df[news_df["title"].notna()]
def prepare_data(df, test_size=0.2, include_title=True, include_author=True):
texts = []
labels = []
for i in range(len(df)):
text = df["text"].iloc[i]
label = df["label"].iloc[i]
if include_title:
text = df["title"].iloc[i] + " - " + text
if include_author:
text = df["author"].iloc[i] + " : " + text
if text and label in [0, 1]:
texts.append(text)
labels.append(label)
return train_test_split(texts, labels, test_size=test_size)
train_texts, valid_texts, train_labels, valid_labels = prepare_data(news_df)
print(len(train_texts), len(train_labels))
print(len(valid_texts), len(valid_labels))
# tokenize the dataset, truncate when passed `max_length`,
# and pad with 0's when less than `max_length`
train_encodings = tokenizer(train_texts, truncation=True, padding=True, max_length=max_length)
valid_encodings = tokenizer(valid_texts, truncation=True, padding=True, max_length=max_length)
class NewsGroupsDataset(torch.utils.data.Dataset):
def __init__(self, encodings, labels):
self.encodings = encodings
self.labels = labels
def __getitem__(self, idx):
item = {k: torch.tensor(v[idx]) for k, v in self.encodings.items()}
item["labels"] = torch.tensor([self.labels[idx]])
return item
def __len__(self):
return len(self.labels)
# convert our tokenized data into a torch Dataset
train_dataset = NewsGroupsDataset(train_encodings, train_labels)
valid_dataset = NewsGroupsDataset(valid_encodings, valid_labels)
# load the model
model = BertForSequenceClassification.from_pretrained(model_name, num_labels=2)
from sklearn.metrics import accuracy_score
def compute_metrics(pred):
labels = pred.label_ids
preds = pred.predictions.argmax(-1)
# calculate accuracy using sklearn's function
acc = accuracy_score(labels, preds)
return {
'accuracy': acc,
}
training_args = TrainingArguments(
output_dir='./results', # output directory
num_train_epochs=1, # total number of training epochs
per_device_train_batch_size=10, # batch size per device during training
per_device_eval_batch_size=20, # batch size for evaluation
warmup_steps=100, # number of warmup steps for learning rate scheduler
logging_dir='./logs', # directory for storing logs
load_best_model_at_end=True, # load the best model when finished training (default metric is loss)
# but you can specify `metric_for_best_model` argument to change to accuracy or other metric
logging_steps=200, # log & save weights each logging_steps
save_steps=200,
evaluation_strategy="steps", # evaluate each `logging_steps`
)
trainer = Trainer(
model=model, # the instantiated Transformers model to be trained
args=training_args, # training arguments, defined above
train_dataset=train_dataset, # training dataset
eval_dataset=valid_dataset, # evaluation dataset
compute_metrics=compute_metrics, # the callback that computes metrics of interest
)
# train the model
trainer.train()
# evaluate the current model after training
trainer.evaluate()
# saving the fine tuned model & tokenizer
model_path = "fake-news-bert-base-uncased"
model.save_pretrained(model_path)
tokenizer.save_pretrained(model_path)
def get_prediction(text, convert_to_label=False):
# prepare our text into tokenized sequence
inputs = tokenizer(text, padding=True, truncation=True, max_length=max_length, return_tensors="pt").to("cuda")
# perform inference to our model
outputs = model(**inputs)
# get output probabilities by doing softmax
probs = outputs[0].softmax(1)
# executing argmax function to get the candidate label
d = {
0: "reliable",
1: "fake"
}
if convert_to_label:
return d[int(probs.argmax())]
else:
return int(probs.argmax())
real_news = """
Tim Tebow Will Attempt Another Comeback, This Time in Baseball - The New York Times",Daniel Victor,"If at first you don’t succeed, try a different sport. Tim Tebow, who was a Heisman quarterback at the University of Florida but was unable to hold an N. F. L. job, is pursuing a career in Major League Baseball. He will hold a workout for M. L. B. teams this month, his agents told ESPN and other news outlets. “This may sound like a publicity stunt, but nothing could be further from the truth,” said Brodie Van Wagenen, of CAA Baseball, part of the sports agency CAA Sports, in the statement. “I have seen Tim’s workouts, and people inside and outside the industry — scouts, executives, players and fans — will be impressed by his talent. ” It’s been over a decade since Tebow, 28, has played baseball full time, which means a comeback would be no easy task. But the former major league catcher Chad Moeller, who said in the statement that he had been training Tebow in Arizona, said he was “beyond impressed with Tim’s athleticism and swing. ” “I see bat speed and power and real baseball talent,” Moeller said. “I truly believe Tim has the skill set and potential to achieve his goal of playing in the major leagues and based on what I have seen over the past two months, it could happen relatively quickly. ” Or, take it from Gary Sheffield, the former outfielder. News of Tebow’s attempted comeback in baseball was greeted with skepticism on Twitter. As a junior at Nease High in Ponte Vedra, Fla. Tebow drew the attention of major league scouts, batting . 494 with four home runs as a left fielder. But he ditched the bat and glove in favor of pigskin, leading Florida to two national championships, in 2007 and 2009. Two former scouts for the Los Angeles Angels told WEEI, a Boston radio station, that Tebow had been under consideration as a high school junior. “’x80’x9cWe wanted to draft him, ’x80’x9cbut he never sent back his information card,” said one of the scouts, Tom Kotchman, referring to a questionnaire the team had sent him. “He had a strong arm and had a lot of power,” said the other scout, Stephen Hargett. “If he would have been there his senior year he definitely would have had a good chance to be drafted. ” “It was just easy for him,” Hargett added. “You thought, If this guy dedicated everything to baseball like he did to football how good could he be?” Tebow’s high school baseball coach, Greg Mullins, told The Sporting News in 2013 that he believed Tebow could have made the major leagues. “He was the leader of the team with his passion, his fire and his energy,” Mullins said. “He loved to play baseball, too. He just had a bigger fire for football. ” Tebow wouldn’t be the first athlete to switch from the N. F. L. to M. L. B. Bo Jackson had one season as a Kansas City Royal, and Deion Sanders played several years for the Atlanta Braves with mixed success. Though Michael Jordan tried to cross over to baseball from basketball as a in 1994, he did not fare as well playing one year for a Chicago White Sox minor league team. As a football player, Tebow was unable to match his college success in the pros. The Denver Broncos drafted him in the first round of the 2010 N. F. L. Draft, and he quickly developed a reputation for clutch performances, including a memorable pass against the Pittsburgh Steelers in the 2011 Wild Card round. But his stats and his passing form weren’t pretty, and he spent just two years in Denver before moving to the Jets in 2012, where he spent his last season on an N. F. L. roster. He was cut during preseason from the New England Patriots in 2013 and from the Philadelphia Eagles in 2015.
"""
get_prediction(real_news, convert_to_label=True)
# read the test set
test_df = pd.read_csv("test.csv")
test_df.head()
# make a copy of the testing set
new_df = test_df.copy()
# add a new column that contains the author, title and article content
new_df["new_text"] = new_df["author"].astype(str) + " : " + new_df["title"].astype(str) + " - " + new_df["text"].astype(str)
new_df.head()
# get the prediction of all the test set
new_df["label"] = new_df["new_text"].apply(get_prediction)
# make the submission file
final_df = new_df[["id", "label"]]
final_df.to_csv("submit_final.csv", index=False)
