Destructive information in text is widespread and dangerous for children and teenagers as well as for adults. Current methods of searching destructive information in text: “keyword search”, “reverse document frequency method” have a number of disadvantages that can cause false positives, which reduces the accuracy of their work. In this article we consider a new developed method of searching destructive information in text, which is used in Python module. This method utilizes Spacy, pymorphy3 libraries which allows us to examine the sentence in detail and delve into its meaning. The developed method reduces false positives and thus increases the efficiency of its use. The paper shows the schemes of sentence parsing, the algorithm of the new method, as well as figures demonstrating its work. The comparative analysis of the new method with analogs is shown.

Keywords: Spacy, disruptive content, information security, TF-IDF, keyword search, pymorphy3, Net Nanny, CyberPatrol, Oculus, child protection