The fake news on social media and various other media is wide spreading and is a mat- ter of serious concern due to its ability to cause a lot of social and national damage with destructive impacts. A lot of research is already focused on detecting it. Here we take three data sets namely ” fake news and real news”, ”ISOT” and ”LIAR”. We try to implement six machine learning models on these data sets and trying to find their accu- racy and precision. The models we uses are Decision Tree, Random Forest, Support vector machine, Naive Bayes, KNN and LSTM. WE use tools like python scikit learn and NLP. Python scikit library can be used for feature extraction and textual analysis. We tries to find out which model works best on which data keeping the complexity of the data in mind. We would like to find a perfect model for any of the regional language. But the constrain is the availability of good dataset . So we try to propose a new dataset.
Table of Contents
1 Introduction
1.1 Literature survey
1.1.1 Novel Stacking Approach for Accurate Detection of Fake News
1.1.2 A New Benchmark Dataset for Fake News Detection
1.1.3 Fake News Detection on Social Media: A Data Mining Perspective
1.1.4 A benchmark study of machine learning models for online fake news detection
1.1.5 Fake News Detection on Social Media using Geometric Deep Learning [1]
1.1.6 CSI: A Hybrid Deep Model for Fake News Detection [2]
1.1.7 TI-CNN: Convolution Neural Networks for Fake News Detection [3]
1.1.8 Fake news detection using deep learning. [4]
1.1.9 A Survey on Natural Language Processing for Fake News Detection
1.1.10 MalayalmFakeNewsDetectionUsingMachine Learning
2 Proposed Architecture
3 Methodology
3.1 Decision Tree
3.2 Random forest
3.3 Support Vector Machine
3.4 Naive Bayes
3.5 KNN
3.6 LSTM
3.7 XGBoosT
4 Implementation
4.1 ISOT
4.2 LIAR
4.3 MALNEWS
5 Results
5.1 Screenshots
5.2 Code:
6 Conclusion
Objectives and Research Themes
The primary objective of this study is to devise and compare machine learning models capable of identifying authentic and fake news articles by analyzing their linguistic structure and auxiliary content. The research investigates the efficacy of supervised learning algorithms across multiple datasets to improve the accuracy of misinformation detection.
- Application of supervised machine learning for fake news classification.
- Comparative performance analysis of six distinct models (Decision Tree, Random Forest, SVM, Naive Bayes, KNN, LSTM).
- Benchmark evaluation using established datasets (ISOT, LIAR) and a custom Malayalam corpus.
- Implementation of Natural Language Processing (NLP) techniques for feature extraction.
Excerpt from the Book
3.3 Support Vector Machine
Support Vector Machines (SVMs) are a powerful machine learning algorithm that can be used for classification, regression, and outlier detection tasks. SVMs are particularly useful when working with high-dimensional datasets, as they can effectively separate data into distinct classes.
SVMs are based on the concept of finding a hyperplane that can separate data into different classes. In a two-dimensional space, the hyperplane is a straight line that separates the data into two classes. In a higher dimensional space, the hyperplane is a subspace that separates the data into multiple classes. SVMs use a process called kernel trick to transform the data into a higher dimensional space where it is easier to separate the data.
The goal of an SVM is to find a hyperplane that maximizes the margin between the classes. The margin is defined as the distance between the hyperplane and the closest data points from each class. The SVM aims to find the hyperplane that maximizes this margin, as it is less likely to overfit and generalize better to new data.
Summary of Chapters
1 Introduction: This chapter provides an overview of fake news dissemination and presents the research goal of using machine learning for automated fact-checking.
2 Proposed Architecture: Describes the design approach used for the experiment, outlining the split between training and testing data and the deployment of statistical packages for feature selection.
3 Methodology: Details the algorithmic approach, including descriptions of the supervised learning models applied in this research such as Decision Tree, Random Forest, SVM, and LSTM.
4 Implementation: Explains the practical application of the study, detailing the datasets utilizados (ISOT, LIAR, MALNEWS) and the technical environment used for classification.
5 Results: Provides illustrative screenshots of the model outputs and presents the source code used for the training and testing of the classifiers.
6 Conclusion: Summarizes the findings of the project, highlighting the performance metrics achieved by the various algorithms on the different datasets.
Keywords
Fake News, Machine Learning, Decision Tree, Random Forest, Support Vector Machine, Naive Bayes, KNN, LSTM, XGBoost, Malayalam Dataset, Natural Language Processing, Classification, Information Retrieval, Data Mining, Supervised Learning
Frequently Asked Questions
What is the core focus of this research?
The research focuses on the automated detection of misinformation through the application of various machine learning models to identify differences between authentic and fake news across multiple platforms.
Which machine learning models are evaluated in the study?
The study evaluates six primary models: Decision Tree, Random Forest, Support Vector Machine (SVM), Naive Bayes, K-Nearest Neighbors (KNN), and Long Short-Term Memory (LSTM) networks.
What is the primary objective or research question?
The study aims to determine which machine learning architectures yield the highest accuracy and precision for fake news detection while considering the inherent complexity and variability of news datasets.
Which scientific methodology is employed?
The work utilizes a quantitative approach involving data collection, pre-processing, feature selection, and the training/testing of classifiers using datasets such as ISOT, LIAR, and a newly proposed Malayalam dataset.
What topics are covered in the main section?
The main sections cover extensive literature surveys on fake news detection techniques, the proposed architectural design for the experiment, detailed explanations of the algorithms, and code implementations for each model.
Which keywords best characterize this work?
Keywords include Fake News, Machine Learning, SVM, LSTM, Random Forest, and Data Mining, reflecting the core technical and thematic components of the research.
How does the study approach the challenge of non-English fake news?
The research addresses language barriers specifically by proposing and implementing a new Malayalam dataset (MALNEWS) and utilizing NLP tools for extraction.
What is the significance of the LIAR dataset mentioned in the study?
The LIAR dataset is highlighted as a large-scale, decade-long collection of manually labeled short statements from POLITIFACT.COM, which allows for more robust research into automated fact-checking.
- Citar trabajo
- Balamurugan Rengeswaran (Autor), Vidhya VP (Autor), 2024, A Study of Tackling Fake News with Machine Learning Approaches, Múnich, GRIN Verlag, https://www.grin.com/document/1471653
