DETECTION OF FAKE ACCOUNTS IN SOCIAL MEDIA

Social media is becoming increasingly used as a source of information, including events during warfare. The fake accounts the social are often for variety cyber-attacks, information-psychological operations, and opinion manipulating analysis online research methods carried out, main metrics attributes in Facebook investigated. Each metric assigned to appropriate categories convenience their gets certain number points depending on conditions from 0 3, which indicate how much every influenced conclusion about fakeness account. levels influence have following meanings: – no influence, 1 weak 2 significant 3 critical influence. For example, if histogram feature reaches level this means that parameter characterizing has impact account fakeness. Otherwise, column at or level, inherent real Thus, based each parameters, we conclude reality analyzed: likes, friends, posts statuses, personal information user photos, considering possible parameters status analysis. A decision-making system supported vector machine developed 9 inputs single output. series experimental was conducted where analyzing well extracting selection realized Facebook. classifier accuracy detection 97% with special prepared dataset parameters.