Fake News Detection Using Various Machine Learning Algorithms

Abstract

Reports have been around for a long time and have a track record of producing accurate results. The fast development of online media stages has increased the adverse consequence of bits of gossip; it accordingly gets essential to early identify them. Numerous techniques have been acquainted with distinguishing reports utilizing the substance or the social setting of information. In any case, most existing strategies overlook or don't investigate viably the engendering example of information in online media, including the grouping of collaborations of web-based media clients with news across time. In this paper, we present a new conversation identification method based on deep learning. By understanding the customers' depiction and the worldwide interconnection of clients' responses, our method makes advantage of the news dissemination dynamics. We target furnishing clients with a stage to check the bits of gossip they hear and counterfeit news that courses through online media stages like Twitter and Facebook. The clients can gather validated realities and check for the wellsprings of the news occasions continuously. Official groups can use the constant stream elements to follow an occasion and individuals engaged with the stream. Tests conducted on Twitter and Weibo datasets indicate that the suggested approach is cutting-edge in terms of performance. Keywords: Propagation Dynamics, User Representation, Data Sets, Counterfeit

Comparison of the Performance of Six Machine Learning Algorithms for Fake News

“…In paper [3], the researchers believe that many fake news stories contain many facts, and conversely, many true news stories contain a lot of false information. Therefore, the researchers evaluated the performance of four machine learning algorithms that work on detecting fake news with the help of natural language processing methods, and after training, the results were as shown in the In research papers [6] and [7], the focus was on machine learning algorithms such as Passive Aggressive classifier and Logistic regression that were applied in a set of algorithms to determine the best algorithm for detecting fake news. In the paper [6], four machine learning algorithms were used, including the Naïve Bayes algorithm, Support vector machine classification, Logistic regression, Passive Aggressive classifier, and some natural language processing techniques such as a bag of words and TD-IDF.…”

“…The practical application results showed that the Passive Aggressive classifier gave the highest accuracy rate of 99.5%. In the paper [7], the researchers sought to determine whether the idea of using artificial intelligence to solve the problem of fake news is valid or not by creating a website that can help users verify fake news. Three machine learning algorithms were used, namely Logistic Regression, Naïve Bayes, and Random Forest, as well as some natural language processing techniques, and the results are shown in the table below: In the paper [8] three models were utilized, including RoBERTa, a pre-trained language model from the BERT family, Bi-LSTM, a deep learning algorithm, and Passive Aggressive Classifier, a machine learning technique.…”

Hindi Fake News Detector

“…Data is collected either by web scraping [1], a website can be used to scrap fake news and true news and various classifiers like SVM [1] [7,13], random forests [1,3,4,8] or the combination of both like in [13] where SVM-RF produces a higher accuracy can be used. Kaggle and other websites can be used to collect datasets but, in this project, web scraping is done.…”

“…Graph Transformer network (GTN) [2] can also be used but the limitation here is emotions and sentiment analysis has not been taken into consideration which vastly effects the results. TF-IDF and count vectorizer [3,5,7,8] is used for feature extraction and a weighing scheme [7] that assigns weights to the features involved can be implemented for feature extraction. A BERT based deep learning approach [10] is implemented and python libraries are used for the same.…”

“…BERT is a pre-trained model that helps in feature extraction, and it is a deep learning model in which each output data is connected to each input. Instead of using regex [13] for pre-processing, this project uses tokenization and generalization [3,5] using Bert Tokenizer for assigning numbers to each token. After the dataset is trained and tested, a confusion matrix [9,6] is implemented to obtain the values of accuracy, precision, recall [2,9,10] which will be the result.…”