GHM edia
B usiness
Article
Global
MDA
Fake News Detection Using
Deep Learning
In today’s world, with the increase in spread of fake news
from social media and various other sources it is becoming
very important to be able to categorize between real news
and fake news. Fake news is a major factor in inciting
riots, carnages, mob lynchings and other social-economic
disturbances. To tackle this problem we propose a model
which detects fake news with the help of Deep Learning and
Natural Language Processing. A Deep Neural Network on self
scraped data set is trained and by using Natural Language
Processing the correlation of words in respective documents
is found and these correlations serve as initial weights for the
deep neural network which predicts a binary label to detect
whether the news is fake or not.
Keyword: Fake News, Safety, Awareness, Natural Language
Processing, Artifical Intelligence, Machine Learing, Deep
Learning, Web Scraping.
Fake news is a false piece of information. Any piece of Fake
news can be created to deliberately misinform or deceive
readers, promote a biased point of view, particular cause or
agenda, and even for the entertainment. Fake news can be
promoted by unauthenticated user ID, social media, printing
of fake news in newspapers maybe due to political pressure
and many more. Spreading of fake news can cause discontent
among people, riots, and even cause loss of trust between
two people and even nations. It also has the potential to
manipulate public thinking in a completely different way.
News and media coverage gets hugely distorted due to
initialization and spread of fake news. Where news can be a
boon, fake news is a bane to the society. These days there are
many anti-social elements, who may instigate and propagate
fake news that may cause societal instability in many layers of
socio-economic-political-cultural aspects. Many instances of
religion clash have been attributed to such fake news.
However, the distinction between a genuine news item and a
fake one is very difficult. Its study hence plays a very important
role in today’s standing. Stopping the propagation of fake
news has taken a huge leap due to many advertisements
by the Government of India. However, its identification
still remains a task filled with discrepancies and many
dependencies. Facebook received more engagement of top
42 | january-february 2019 |
20 fake news stories about 2016 U.S. presidential elections
than top 20 election stories from 19 major media outlets
(source: Buzzfeed). The most difficult part comes from the
fact that even human eye cannot distinguish perfectly
between real and fake news. One study found that 75% of the
time people find fake news somewhat or very accurate. Fake
news on social media has experienced a upturn due to the
increased concern on the spread and negative effects of fake
news. In the previous year, in January 2017, a German official
even accepted the dilemma and referring to the spread of
fake newsstated that “they are dealing with a phenomenon
of dimension never seen before” referring to the spread of
fake news. So, detection of fake news becomes most crucial
part we need to do in order to have sorted lives.
Machine Learning tackles this problem with the help of Deep
Learning (which constitutes imitation of a human brain and
consistes Artificial Neural Networks which imitate neurons
in a brain) and Natural Language Processing (which helps
the machine (artificial brain) to make sense of textual data).
Let us refer this algorithmic model as artificial brain for now.
Any news article whose validity is doubted can be fed into
this “artificial brain” and it predicts whether it is fake or
real. The magic is made possible as the artificial brain learns
to differentiate fake news from real with the help of past
experiences (thus the machine is learning).
This artificial brain learns by processing examples of how the
real and fake news look like and learns to differetiate between
them. This artificial brain breaks the article into segments
which relates words according to context and checks the
validity of these correlations (example: football is related to
Lionel Messi and cricket to Sachin Tendulkar, not the other
way around! But all these words are related to sports).
These correlations are handled by advanced mathematical
operations, equations and algorithms and are converted
to numeric data from which our artificial brain learns and
becomes Artificially Intelligent thus being able to predict the
validity of a news.
Abhishek Verma
Vanshika Mittal
Global MDA Journal
Journal