International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 12 Issue: 03 | Mar 2025
p-ISSN: 2395-0072
www.irjet.net
Detection of Fake and Real Messages using Machine Learning Techniques Akshatha T, Akshatha S.A 1Professor, Dept. of Artificial Intelligence and Data Science Engineering, Bearys Institute of Technology,
Mangalore, Karnataka
2Professor, Dept. of Computer Science and Engineering, Bearys Institute of Technology, Mangalore, Karnataka
---------------------------------------------------------------------***--------------------------------------------------------------------Processing (NLP) examines text structure, sentiment, and Abstract - The objective of this paper is to identify real or
word patterns to detect inconsistencies. Deep learning models analyze large datasets to recognize characteristics common in fake news. These automated systems enhance the speed and accuracy of misinformation detection compared to manual methods.
fake message by using machine learning. Machine learning algorithms play a crucial role in differentiating between authentic and false news articles. Here we have used Logistic Regression, Gradient Boosting Classifier (GBC), and Random Forest Classifier because of their effectiveness in handling text-based classification tasks. Two datasets containing 1,000 real and fake messages are preprocessed and then split into training and testing datasets. The testing dataset is used to evaluate the trained models. After four models were trained classification report was generated and manual test was performed to check whether massage is real of fake.
c. Metadata and Source Verification Evaluating the source of a message is essential in determining its authenticity. Trusted sources are typically transparent about their authorship and maintain a strong track record of accuracy. Examining metadata, including publication date, author details, and domain credibility, helps in assessing reliability. Websites that provide little information about their origins or frequently publish exaggerated content may be unreliable.
Key Words: Machine learning, fake and real news, Logistic Regression, Gradient Boosting Classifier (GBC), Random Forest Classifier.
1. INTRODUCTION
d. User Behavior Analysis
Detecting real and fake news is crucial to preventing misinformation, safeguarding democracy, and minimizing negative societal impacts. False information can deceive people, influence opinions, and cause unnecessary fear, particularly in areas like politics, health, and finance. It also erodes trust in media and institutions while enabling fraud and cybercrimes. By distinguishing between genuine and misleading news, we can ensure the spread of accurate information, promote critical thinking, and help individuals make well-informed decisions based on credible sources. There exist various methods of detecting fake and original news.
The spread of misinformation is often linked to bots and coordinated campaigns. By analyzing social media activity, unusual trends can be detected, such as new accounts repeatedly posting the same message. Another sign of misinformation is abnormal engagement, where content receives many shares but limited meaningful discussion. Identifying these patterns helps in recognizing and limiting the spread of fake messages. e. Image and Video Forensics Misleading messages frequently contain altered images or videos to distort facts. Reverse image search tools, like Google Reverse Image Search and TinEye, help verify whether an image has been edited or used out of context. Additionally, deepfake detection technology examines inconsistencies in video elements, such as unnatural facial movements or lighting, which may indicate digital manipulation.
a. Manual Fact-Checking One of the most effective ways to verify information is through manual fact-checking. Websites like Snopes, PolitiFact, and FactCheck.org analyze claims using credible sources and expert evaluations. Additionally, individuals can compare the information with well-established news sources to determine its reliability. If a message lacks confirmation from trusted organizations, it is likely misleading or false.
In our research we have used machine learning methods to identify the fake and original news. The methodology involves first collecting the data, classifying the data, data preprocessing, feature extraction, Model training and Evaluation, Model testing and then result.
b. AI and Machine Learning-Based Detection Artificial intelligence (AI) and machine learning (ML) play a key role in identifying misinformation. Natural Language
© 2025, IRJET
|
Impact Factor value: 8.315
|
ISO 9001:2008 Certified Journal
|
Page 909