FAKE NEWS DETECTION
About the Project
The Fake News Detection project addresses the proliferation of misinformation in Urdu news. Utilizing a custom Naive Bayes model, this project provides an automated solution to identify and flag potentially fake news articles, helping Urdu-speaking users distinguish credible information from falsehoods.
Technologies: React JS, Node JS, Python, FASTAPI, OpenCV, Pytorch
Services: Web Development, AI Integration
Our Goal
The primary goal of the Fake News Detection project is to combat the spread of misinformation in Urdu news by developing an automated solution that can accurately identify and flag potentially fake news articles. This initiative aims to empower Urdu-speaking users with the tools to discern credible information from falsehoods, thereby fostering a more informed and aware society.
Challenges
Identifying fake news in Urdu, considering linguistic nuances and limited availability of labeled datasets.
Processing and analyzing text written in Urdu script.
Ensuring high accuracy in detection to maintain user trust.
Continuously updating the model to adapt to new patterns in fake news.
Solutions
Data Collection and Preprocessing: Collected a dataset of Urdu news articles from multiple sources, including labeled datasets of known fake and real news. Preprocessing involved text normalization, tokenization, and handling specific challenges related to Urdu script.
Feature Extraction: Utilized techniques such as TF-IDF to convert text into numerical features that the Naive Bayes model could process effectively.
Model Development: Developed a custom Naive Bayes model tailored for the Urdu language. The model was trained to classify news articles as fake or real based on extracted features.
Deployment: Integrated the model into a user-friendly web application, allowing users to input Urdu news articles or URLs and receive real-time feedback on the likelihood of the news being fake.
Model Evaluation: Evaluated the model using metrics such as accuracy, precision, recall, and F1-score. The custom Naive Bayes model demonstrated strong performance due to its simplicity and effectiveness for text classification tasks.
Impacts
- Empowered Urdu-speaking individuals and organizations to combat misinformation.
- Contributed to a more informed and aware society within the Urdu-speaking community.
- Provided a scalable solution that can be integrated into various platforms, including social media, news websites, and fact-checking organizations.
Results
Achieved a detection accuracy of 90%, significantly aiding in the identification of fake news in Urdu.
Enabled users to make informed decisions by providing clear, actionable insights.
Enhanced user trust in the platform by maintaining high accuracy and continuously updating the model.