You are on page 1of 4

Advanced Fake News Detection

Problem Statement Title: Advanced Fake News Detection System


Problem Statement Description:
Design and develop a technological solution/software tool for Tracking &
Tracing Fake News and its origin using official sources as the input filter. The
solution should have a mechanism to mitigate the impact of the spread of Fake
News by auto-populating the fake news spreaders’ inboxes with the
official/authenticated news content.
Introduction:
The Internet is one of the important inventions and different social media
platforms are accessible to everyone. Any user can make a post or spread the
news that these platforms do not verify. Some users try to spread fake news
which can be propaganda against an individual, society, or political party. It is
difficult to detect fake news by humans. So, there is a need for machine learning,
a part of artificial intelligence, that can detect fake news automatically.
The solution can be a web browser plugin that uses machine learning
algorithms to identify fake news. The plugin can be installed on the user’s
browser and will scan all incoming media, including images, videos, and audio,
for fake news and if any fake news is detected it will send the user an alert
message.
Idea Details:
The world is changing rapidly. We have several advantages and
disadvantages in this digital world and fake news is one of the disadvantages.
Someone can easily spread fake news which can harm the reputation of a person
or an organization.
Machine learning is used most of the time for prediction purposes or to
detect something that is hidden. A variety of algorithms are available that include
supervised, unsupervised, and reinforcement machine learning algorithms which
have to be trained first with a dataset. After training, these algorithms can be used
to perform different tasks.
Methodology:
The processes that are involved in the project being employed here are:
1. Data Collection
2. Machine Learning
3. Detection of Fake News
4. Tracing the Origin
5. Mitigating the impact
6. Alert and reporting
7. Continuous Improvement

1. Data Collection:
The solution would collect data from various official news sources
such as government websites, news agencies, and other authenticated
sources. This data would be used as the baseline for authentic news
content.
2. Machine Learning:
The solution would analyze the collected data using machine
learning algorithms and create a baseline model for authentic news
content. This model would be used to detect fake news and its origin.
3. Detection of Fake News:
Once the baseline model is established, the solution would start
monitoring online content, including social media platforms, news
websites, and other sources, to detect any fake news that is spreading.
4. Tracing the Origin:
Once the fake news is detected, the solution would use machine
learning to trace its origin by analyzing its spread pattern across various
social media platforms, websites, and other sources. This would help
identify the source of the fake news and track it to its origin.
5. Mitigating the Impact:
Once the source of the fake news is identified, the solution would
automatically send authentic news content to the fake news spreaders’
inboxes using machine learning. This would help mitigate the impact of
the spread of fake news.
6. Alert and Reporting:
The solution would also generate alerts and reports to relevant
authorities and official news agencies, including the source of the fake
news, the extent of its spread, and any potential impact on society.
7. Continuous Improvement:
The solution would continuously learn and improve by analyzing
new data and updating its baseline model for authentic news content to
ensure the highest accuracy in detecting and mitigating the impact of
fake news.
Several Technologies can be used to build a fake news-detecting machine
learning model application. Some of them include:
1. Natural Language Processing (NLP)
2. Deep Learning
3. Python Programming Language
4. Web Scraping
5. Data Visualization
Which Machine Learning supervised classifiers can be used for detecting
fake news?
Some of the popular classifiers are:
1. Support Vector Machine (SVM)
2. Naïve Bayes
3. Logistic Regression
4. Random Forests
5. Recurrent Neural Network
6. K-Nearest Neighbor
7. Decision Tree
How machine learning classifiers are trained for detecting fake news?
Training the classifiers of machine learning is an important task. This plays
an important role in the accuracy of the results of these classifiers.
Dataset

Removing Stop Words for Stemming

Test dataset & Train dataset

Training Classifier

Experimental Results
Fig: Training Dataset
Conclusion:
A Machine Learning Model is built which can detect Fake News and is
embedded into an application/website to alert the user.

You might also like