On cloud computing systems, machine learning techniques are used to detect fake news.
Shalini kumari
Dr. Priti bihade
Dr.Pravin Kulurkar
G H Raisoni college of engineering and management
Abstract :The exponential growth of information shared on the internet, particularly through social media platforms, has made distinguishing between authentic and fake news increasingly challenging. With the proliferation of web-based networking media, a significant portion of smartphone users now prefer reading news on social media rather than traditional websites. However, the authenticity of information published on these platforms often remains unverified, leading to the rapid dissemination of misinformation.
This ease of sharing has exacerbated the problem, contributing to the exponential spread of fake news. As a result, fake news has emerged as a critical issue, especially with the internet's widespread accessibility and its pivotal role in shaping public opinion. Addressing this challenge requires robust mechanisms to categorize news as either legitimate or illegitimate.
To tackle this issue, we developed a framework leveraging various machine learning (ML) techniques. Python, chosen for its versatility and extensive libraries, served as the primary scripting language for implementation. The framework employs several ML methods, including K-Nearest Neighbors (KNN) and Decision Trees (DT), complemented by an integrated approach using advanced ensemble techniques such as Random Forest (RF), Gradient Boosting (GB), and custom ensemble methods. These custom methods, including Stacking and Maximum Voting Classifiers, demonstrated superior performance in identifying fake news.
Notably, the Stacking approach, combining classifiers like KNN, Support Vector Classifier (SVC), and Logistic Regression (LR) in a custom ensemble, achieved the highest accuracy in categorizing news. This integrated methodology underscores the potential of combining multiple ML techniques to enhance the efficiency and reliability of fake news detection systems.
Keywords: Natural Language Processing (NLP), Natural Language Toolkit (NLTK), Term Frequency-Inverse Document Frequency (tf-idf) Vectorizer, Ln-built and Custom ensembled Machine Learning (ML) Models, Support Vector Classifier (SVC), Logistic Regression (LR), K- Nearest Neighbors (KNN)
