Open AccessFAKE NEWS DETECTION SYSTEM BASED ON DATA SCIENCE
Author(s) -
Volodymyr Bazylevych,
Maria Prybytko
Publication year - 2020
Publication title -
tehnìčnì nauki ta tehnologìï
Language(s) - English
Resource type - Journals
eISSN - 2519-4569
pISSN - 2411-5363
DOI - 10.25140/2411-5363-2020-4(22)-91-95
Subject(s) - computer science , data science , random forest , classifier (uml) , data mining , decision tree , fake news , information retrieval , artificial intelligence , internet privacy
Urgency of the research. Today, the task of analyzing the veracity of information in the news, which filled all existing channels for obtaining information, is relevant. Its urgency is related to the need to prevent panic by obtaining inaccurate information, debunking pseudo-scientific facts that can threaten people's lives, combating political propaganda and others.Target settingThis article focuses on the concept of developing a system for detecting fake news, analysis of existing systems and their principles of operation, principles of construction of their algorithms and features of their use.Actual scientific researches and issues analysis.Recent open publications, statistics, and corporate reports were reviewed.Uninvestigated parts of general matters defining.File analysis will be performed using three methods / classifiers and without the use of PassiveAgressive classifier. The calculation and derivation of results is performed by constructing error matrices and calculating accuracy.The research objective.The main purpose of the work is to create a system for detecting fake news on the basis of the considered materials and to achieve the highest possible accuracy.Presenting main material. Input data for the study were selected, prepared and analyzed. Data were studied using the meth-ods /classifiers of Logistic Regression, Decision Tree and Random Forest. The accuracy of detecting fake news is calculated.Conclusions.The proposed system allows to classify news as “fake”or “true ”with an accuracy of 98-99%