Akhond, Mostafijur RahmanSaif, MehruzKanon, MD. Kamal HaqueHasan, NazmulHossen, MD. ShamimAnannya, Fatema Zohra2021-10-102021-10-1020212021-06ID 19101665ID 19201139ID 19301277ID 15301101ID 17101176http://hdl.handle.net/10361/15199Cataloged from PDF version of thesis.Includes bibliographical references (page 41-42).This thesis is submitted in partial fulfillment of the requirements for the degree of Bachelor of Science in Computer Science and Engineering, 2021.The World Wide Web’s launch and the rapid adoption of social media platforms (such as Facebook and Twitter) paved the way for unparalleled levels of information diffusion in human history. Consumers are creating and sharing more information on social media platforms than ever before, some of it is erroneous, deceptive, or has no influence on reality. Access to news information has become considerably simpler and more comfortable thanks to the Internet and social media. Online users may often follow events of interest, and the widespread usage of mobile devices makes this process easier. However, with great potential comes enormous responsibility. There are also a number of websites dedicated nearly entirely to the dissemination of fake news. Since it’s a serious issue with a large-scale dataset, identification of fake news is very vital in this era, as social media and online newspapers are in large numbers in the web arena. That’s why it is easy to spread rumors and create chaos. Also, the size of data sets is increasing day by day. Data is expanding at a quicker rate than processing rates. As a result, algorithms that need a huge quantity of data and processing are frequently conducted on a distributed computing system that separates multiple nodes on several machines which have concurrency of components and lack of a global clock. Also, nobody has used a distributed system to detect fake news before. In our paper, we tried to run 4 PySpark algorithms based on SPARK-Context which provides massive storage for big data processing and analysis and also has been found to be 100 times quicker in-memory, while disk performance was shown to be 10 times quicker on several devices at the same time. So that we can control and real-time monitoring over the news and data before it goes viral in the media.42 pagesenBrac University theses are protected by copyright. They may be viewed from this source for any purpose, but reproduction or distribution in any format is prohibited without written permission.PySpark MLRDD(Resilient Distributed Dataset)Random ForestFactorization Machine ClassifierLinear SVCLogistic RegressionMachine LearningThesis