The article is devoted to the methods of identifying the user's of social networks psychological features and construction on these features user's vulnerabilities profile on the basis of the user's audio preferenses on his personal page in social network VK.com. The description of the software that implements these methods is illustrated in this article. The development of the methodology is based on the generalization of studies that are focused on identifying the correlation between user's musical preferences and his psychological features. The implementation of this method and the results of audio recognition software testing are also described in this article.